Skip to main content

Full text of "The Philippine journal of science"

See other formats










ISflO 



&■■»** 



THE PHIUPPINE 

JOURNAL OF SCIENCE 

EDITED BY 

PAUL C. FREER, M. D., Ph. D. 

WITH THE COOPERATION OF 

GEORGE F. RICHMOND, M. S.; W. D. SMITH, Ph. D. 

ALVIN J. COX, Ph. D.; GEORGE I. ADAMS, D. Sc 

RAYMOND E. BACON, Ph. D. ; H. D. GIBBS, B. S. 

NELLIE LOUISE COOK, B. L., M. A. 

PUBLISHED BY 

THE BUREAU OF SCIENCE 

OF THE 

GOVERNMENT OF THE PHILIPPINE ISLANDS 

A. Chemical and Geological Sciences and 
The Industries 

Volume V 

1910 

With 31 Plates, 43 Figures, 2 Maps, and 1 Diagram 




MANILA 

BUREAU OF PRINTING 

1910 



CONTENTS. 



No. 1, January, 1910. 

I, Richmond, George F. Tayabas Petroleum 1 

II. GiBBS, H. D. The Compounds which Cause the Red Coloration 
of Aniline: I. The Effect of Oxygen and Ozone and the 
Influence of Light in the Presence of Oxygen 9 

III. HuLBURT, Clinton H., and Bacon, R. F. A Liquid Concave 

Mirror 19 

IV. Holmes, W. C. Distilled Liquors: Whisky (Rye, Bourbon and 

Scotch), Brandy (Cognac) and Gim 23 

V. Vivencio del Rosario, Mariano. On the Determination of 

Aldehydes in Distilled Liquors 29 

VI. GiBBS, H. D., and Agcaoili, F. Lard from Wild and Domestic 
Philippine Hogs and the Changes in the Constants Produced 

by Feeding Copra Cake 33 

VIL Editorial 45 

VIIL Reviews 55 

No. 2, March, 1910. 

IX. Adams, George I. Geological Reconnaissance of Southwestern 

Luzon ^'^ 

X. Reibling, W. C. Concrete Construction in Manila and the 

Philippine Islands 117 

XI. Adams, George I., and Pratt, Wallace E. Philippine Pot- 
tery 143 

XII. Iddings, Joseph P. The Petrography of some Igneous Rocks 

of the Philippines 1^^ 

No. 3, May, 1910. 

XIIL Freer, Paul C. The Study of Manila Copal 171 

XIV. FoxwoRTHY, F. W. The Almaciga Tree: Agathis Alba (Lam.) .. 173 

XV. Richmond, George F. Manila Copal 177 

XVI. Brooks, Benjamin T. The Destructive Distillation of Manila 

Copal 

XVII. Brooks, Benjamin T. The Oxidation of Manila Copal by the 

Air ' 

XVIII. Brooks, Benjamin T. The Oleoresins of Pinus Insularis Endl .... 229 

in 



203 
219 



>r^i:i 52 



IV CONTENTS. 

No. 4, July, 1910. 

XIX. Richmond, George F. Philippine Fibers and Fibrous Substances : 

Their Suitability for Paper Making 233 

XX. Bacon, Raymond F. Philippine Terpenes and Essential Oils, 

IV 257 

XXI. Bacon, Raymond F. A Preliminary Study of the Effect of 
Tropical Sunlight on the Atmosphere, with Some Notes on 

Radioactive Phenomena in the Philippines 267 

XXII. Bacon, Raymond F. A Solution of Oxalic Acid and Uranium 

Salts as a Chemical Photometer 281 

XXIII. Reviews 305 

Noi 6, November, 1910. 

XXIV. Smith, Warren D. The Essential ^Features of the Geology of 

the Philippine Islands 307 

XXV. Smith, Warren D. Geologic Reconnaissance of Mindanao and 

Sulu: II. Physiography 345 

XXVI. Reviews ^ : 365 

No. 6, December, 1910. 

XXVII. Reibling, W. C, and Reyes, F. D. Physical and Chemical 

Properties of Portland Cement 367 

XXVIII.. GiBBS, H. D. The Compounds which Cause the Red Coloration 
of Aniline: II. The llffect of Sunlight in the Absence of 
Oxygen and Oxidizing Influences and a Comparison with the 

Behavior of Mono- and Dimethylaniline 419 

XXIX. Brooks, Benjamin T. The Natural Dyes and Coloring Matters 

of the Philippines 439 

XXX. Index, Title-page and Contents 453 



JUST ISSUED. 

Order No. 418. 

THE SUGAR INDUSTRY IN THE ISLAND OF NEGROS 

By HERBERT S. WALKER. 

145 pages, 10 photographic plates, and 1 map. 

Price $1.25, United States currency, postpaid. 



CONTENTS. 



Page. 

Introduction 9 

General information regarding Negros 10 

Geographical location 10 

Size, shape, and area 10 

Mountains 11 

Rivers 11 

Climate 11 

The sugar belt 14 

The east coast 16 

Other sugar-producing districts of Negros.. 15 

History of sugar production in Negros 16 

Recent statistics 16 

Varieties of cane grown in Negros 1 18 

Cane diseases and insect enemies 19 

Nationality of the planters ^ 19 

Native labor: Difficulties, past and present 20 

The principal sugar-producing districts of Negros 22 

Silay 22 

Bago 27 

Pontevedra-La Carlota 37 

Binalbagan-Isabela 43 

Ilog-Cabancalan 49 

San Carlos 55 

Bais 64 

The soil of Negros compared with that of other sugar-producing countries.... 68 

Average composition of the soils of Negros .-... 68 

Hawaiian soils 70 

Egyptian soils 70 

Louisiana soils 71 

Java soils 71 

Demerara soils 71 

Berbice soils — .. 72 

Mauritius soils 72 

Comparison of Negros soils with those of other countries 72 

Fertilization in Negros 73 

[OVBB.] 



Order No. 412. 
Page. 

The cane of Negros 76 

Average composition of the purple or native sugar cane in Negros 70 

Other varieties of cane grown in Negros 77 

• Cane in the Hawaiian Islands 79 

Egyptian cane 80 

Java cane 80 

Louisiana cane l 80 

West Indian cane - 81 

Negros as compared with other countries in respect to the quality of 

cane 81 

Desirability of introducing other varieties of cane 81 

The cultivation of sugar cane and the production of sugar as carried on 

at the present time in Negros 82 

Preparation of the soil 82 

Preparation of the seed 83 

Planting 83 

Cultural operations after planting 84 

Cultivation of ratoon canes 85 

Period growth of the cane 86 

Cost of cultivation 87 

Cutting the cane 90 

Transporting the cane to the mill 91 

Cost of cutting the cane and transporting it to the mill 92 

Manufacture of sugar from the cane 92 

' Extraction of the juice - 92 

Manufacture of sugar from the juice 99 

Quality of the sugar produced in Negros 109 

Cost of manufacture 112 

Transportation and sale of the sugar 113 

Estimate of average cost of same 114 

Quantitative experiments to determine the weight of sugar produced 

from a given weight of cane 114 

Mill Control No. 1, hacienda San Jose 117 

Mill Control No. 2, hacienda San Jose 118 

Calculation of the average cost of producing sugar in Negros by the 

methods now employed 123 

Cost for labor alone 123 

Estimate of fixed charges for maintenance and depreciation of plant 

and interest on the capital invested 125 

Total cost of production - 126 

Possibilities for improvement 126 

In cultivation 126 

In manufacture 127 

Advantages of a change to modern methods of manufacture 132 

The future of Negros 1^3 

Summary - 1^4 

Appendix v ;■ ^^'^ 

An Investigation to discover if diseases of the sugar cane exist in 

Negros J^9 

Index 143 

NOTE. 

Orders should be sent to the Business Manager, Philippine Journal of Science, Manila, 
or to any of the below-listed agents. Please give Order Number. 

FOREIGN AGENTS. 

The Macmillan Company, 64-66 Fifth Avenue, New York City, U. S. A. 

Messrs Wm. Wesley & Son, 28 Essex Street, Strand, London, W. C, England. 

Mr. Martinus Nijhoff, Nobelstraat 18, The Hague, Holland. 

Messrs. Mayer & Muller, Prinz Louis Ferdinandstrasse 2, Berlin, N. W., Germany. 

Messrs. Kelley & Walsh, Limited. 32 Raffles Place, Singapore, Straits Settlements. 

Messrs. A. M. & J. Ferguson, 10 Baillie Street, Colombo, Ceylon. 

CIRCULARS AND DESCRIPTIVE MATTER SENT ON APPLICATION. 



-p THE PHILIPPINE 



Journal of Science 

A. Chemical and Geological Science 
AND THE Industries 



Vol. V JANUARY, 1910 No. 1 



TAYABAS PETROLEUM. 



By George F. Richmond. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. /.) 



The first examination of a specimen of crude petroleum reported to 
have been taken from a well in Tayabas Province was made in February, 
1908, when approximately 6 liters of oil were submitted by Castle 
Brothers-Wolf and Sons of Manila. 

The following report^, Laboratory No. 56695, was rendered by the 
Bureau of Science. 

REPORT ON CRUDE PETROLEUM. 

The crude oil as received was especially mobile, brown in color, fluorescent and 
contained no water or sediment. Its specific gravity was 0.826 at 15" C, which 
is somewhat higher than most American petroleums, but decidedly lower than 
European crude oils. 

As it was so clean and dry, no preliminary purification was made, but it was 
distilled directly into three portions. The first fraction coming over from the 
point at which the crude oil begins to boil — viz, fibout 80° up to 150* — comprised 
35 to 40 per cent of the original oil by volume. The middle fraction was collected 
between 150° and 300° and amounted to from 40 to 45 per cent; the residue 
left in the still above 300° measured 15 to 20 per cent. 

Each of the above fractions was in turn divided into three portions by fractional 
distillation. Of the naphtha distillate, 50 per cent boiled between 80° and 110° 
and is essentially naphtha; 38 per cent boiled between 110° and 150° and consists 
mainly of ligro'm and benzine; 11 per cent remained in the still above 150° and 
was added to the burning-oil distillate. 

The burning-oil distillate lost 3.26 per cent below 150°, which was returned 
to the benzine fraction; 90.33 per cent boiled between 150° and 300° and consists of 
almost water-white kerosene; 5.25 per cent remained in the still above 300°, and 
hence was added to the residuum. 
93401 



2 RICHMOND. 

The residuum contained 13.36 per cent of burning oil below 300°; 72.27 per 
cent lubricating oils boiling between 300° and 400°; and 14.37 per cent thick 
residuum, consisting mainly of paraffine scale and coke. 

The interesting features in connection with the composition of this crude 
petroleum are the high percentages of light oils and the comparatively low yield 
of heavy lubricants. 

The specific gravities, boiling points and percentages of the various fractions 
are given in the following table: 

Crude petroleum distilled. 
(Sp. Gr. 0.826 at 15° C.) 



Naphtha distillate below 150°, 
35 to 40 per cent. 



Burning oil distillate 150° to 
300° C, 40 to 45 per cent. 



Residuum above 300° C, 15 
to 20 per cent. 



Naphtha 80° 

to 110°, 50 per 

cent; sp. gr. 

0. 746. 



Above 150°, 
11 per cent. 



Below 150° 
3.26 per cent. 



Benzine 110°-150°, 

38 per cent; sp. gr. 

0.775. 



Above 300°, 
5.25 per cent. 



300-400° C, 

72.27 per 

cent. 



Kerosene 150°-300°,' 

90.33 per cent; sp. 

gr. 0.824. 



Below 
300° C, 
13.36 per 
cent; sp. 
gr. 0.876. 



Residue, par- 

affine coke, 

14.37 per 

cent. 



(Signed) G. F. Richmond. 

In Aprils 1909, the same sample of crude petroleum was further 
examined upon request of the Standard Oil Company of New York, 
and the following report, Laboratory No. 68192, was rendered : 

Origin. — Said to have been collected from a well in Tayabas Province, Luzon, 
Philippine Islands. 

Appearance. — The crude oil as received was especially mobile, free from water 
or sediment, slightly brownish, wine-red by transmitted light, and with a deep 
blue fluorescence when viewed by reflected light. It possessed no objectionable 
odor although the smell was strongly suggestive of light oils. 

The amount of sample at our disposal — viz, 250 cubic centimeters — did not 
admit of a complete examination, hence the presence or absence of sulphur or 
the exact nature of the bases and amount of solid hydrocarbons present have not 
been determined. 

The following tests were applied with a view of detecting rny admixture of 
distilled products or residue. 

According to the New York Produce Exchange, crude oil in the United States 
shall be understood to be pure, natural oil, neither steamed nor treated, free from 
water, sediment or any adulteration, of the gravity 43° to 48° Baume (Q.809 to 
0.786). The rule regarding specific gravity has since been relaxed because much 
crude petroleum now obtained has a greater density than 43° Baum6. 

In order to determine whether a petroleoum is a "pure natural oil,'^ a sample 
is subjected to fractional distillation, each fraction being one-tenth of the crude oil 
by volume and the densities of the several distillates are determined. A regular 
gradation in the densities of the fractions is regarded as a satisfactory indication 
that the oil is a natural product.^ 



^Redwood. Petroleum and its Products (1906), 2 2d ed., 531. 



TAYABAS PETROLEUM. 3 

EXPEBTMENT 1. 

One hundred cubic centimeters of the crude oil in question when subjected 
to the above treatment exhibited the following properties: 



Fraction. 


Specific 
gravity 
at30oc. 


Baum6. 


Dlflfer- 
ence. 


First tenth _. 

Second tenth 


0.728 
0.747 
0.759 
0.773 
0.792 
0.815 
0.841 
0.865 


63° 
58^' 
55° 
610 
470 

42° 
370 
32° 


5 
3 
4 

4 
5 
5 
6 


Third tenth _ ._ 


Fourth tenth _ __ 


Fifth tenth 


Sixth tenth 


Seventh tenth 


Eighth tenth _ _ _. 





It will be noted that the gradation in the densities, as indicated in the column 
of differences, is very regular and not indicative of any admixture of distilled 
products. 

Experiment 2. 

Another 100 cubic centimeters of the crude oil was fractioned according to 
Engler's Method." 

All the details of the distillation including the exact requirements of apparatus 
as outlined, were closely followed. The fractions were collected at 25° C. intervals 
with results as follows: 



Fraction. 


I 

80-100 5.7 
100-125 \ 'll. 3 


1 _ , .. 


2 _ 


3 _ - _ 


125-150 : 22.0 






Total amount below 150° 


! 39.0 






4 


1 
150-175 ! 10.0 


5 


175-200 : 7.5 


6_ _ _ _ 


200-225 7.6 
225-250 ; 9.2 
250-275 ! 6.9 
275-300 ' 6.0 


7 ._ 


8 


9 „ . :. 






Total amount 150°-300° C 


47.1 


10 (total residue above 300° C.) by difference 


__ i 13.9 







It will thus be seen that this oil is unusual in its relatively high proportion 
of light oils and comparatively low percentage of heavy lubricants. 

Of 89 samples of crude oil from different parts of the world, tabulated by 
Redwood,* only two show a higher percentage of light oils, viz, one Russian and 
one Italian oil, which contained 48.9 and 43.9 per cent of light oils, respectively; 
yet the oil examined is not without precedent, for the Sumatra oils as a class show 
a very similar composition. 

2 /bid., 534-536. 
^ Loc. cit. 



4 RICHMOND. 

For the purposes of illustration and comparison the more closely allied oils from 
other sources are tabulated with the oil in question as follows: 



Source. 


Spe- 
cific 
gravity. 


Light oils 
below 
160°C. 


Kero- 
sene, 
150-300°C. 


Lubri- 
cating 
oils. 


Coke. 


Russian 


0.780 
0.787 
0.807 
0.771 
0.813 
0.769 
0.777 
0.826 


Per ct. 
48.9 

43.9 

37.3 

37.2 

30.4 

30.0 

28.0 

39.0 


Per ct. 
43.9 

46.5 

40.4 

52.0 

51.7 

45.0 

49.0 

47.1 


Per ct. 


Per ct. 


Italian 


5.5 
19.9 

5.9 
11.2 
16.1 
14.5 
13.9 


0.0 


Do 


0.3 
0.5 
1.7 
0.0 
0.0 
0.0 


Sumatra 


Do 


Do 


Do 


Tavabas 





The oil examined has nothing in common with oils from other Philippine 
sources examined by Redwood, or in this laboratory. It is also very unlike the 
recorded data on Burma, Borneo, and Java oils. I can not explain its high 
gravity, 0.826 to 0.827, from the findings of the above tests, but there is no 
indication to the undersigned that other crude oil, distilled petroleum products or 
residue have been added. Its extreme clearness and freedom from water and 
sediment strongly suggest that the oil may have been subjected to some form 
of clarification. 

(Signed) G. F. Richmond. 

In July, 1909^ Dr. George I. Adams^ of the division of mines, Bureau 
of Science, visited the Tayabas oil district and collected two authentic 
samples of oil from the well on the Bayhay River for further examination. 

Sample No. 1 was pumped from the well just as it was found, after having 
been undisturbed for some time. 

Sample No. 2 was pumped from the well the following day after it had been 
emptied in securing Sample No. 1. 

Accordingly, sample No. 1 represents the oil after it had lost some of 
its more volatile constituents and sample No. 2 represents it as it comes 
from the oil-bearing strata. A preliminary examination of the two sam- 
ples gave the following results : 





No. 1. 


No. 2. 


Specific gravity of filtered oil at 15°. 5 


0.845 


0.831. 

70° 

36.5 per cent. 

48.75 per cent. 

14.75 per cent. 


Initial boiling point - 


80° 

27 per cent _._ 
66.75 per cent_ 
16.25 percent- 
inn nn - 


1st fraction (light oils) 70 to 150° 


2nd fraction (burning oils) 150 to 300° — _ __ 


Residue above 300°., by difference 


100.00. 





The higher specific gravity, higher initial boiling temperature and 
lower proportion of the light oil fraction of sample No. 1 is readily 
explained by the character of the two, both of which are authentic. The 
chemical and physical constants of sample No. 2 substantiate the authen- 



TAYABAS PETROLEUM. 5 

ticity of the sample of oil (Laboratory Nos. 56695 and 68192) originally 
examined. For this reason and also because it more nearly represents 
the fresh oil from the oil bearing strata, sample No. 2 was selected for 
further study. 

OIL SAMPLE NO. 2. 

Physical and chemical properties. — Color slightly brownish, wine-red 
by transmitted light and bluish flourescent by reflected light. It is 
particularly mobile and free from water and sediment. No deposit of 
solid hydrocarbons occurs when cooled to —15°. The specific gravity 
is 0.8325 (39° Baume) at 15°. Flash point (Abel's closed cup) 0°. 

The crude oil possesses no objectionable odor and gives negative quali- 
tative tests for sulphur. 

Fractional distillation test. — One hundred cubic centimeters of the 
crude oil measured at 15° were transferred to a distilling flask and 
carefully fractioned under the exact conditions prescribed by Engler. 

The initial boiling temperature was taken when the first drop of 
distillate fell from a 75-centimeter condensing* tube of 75° inclination. 
The average results of duplicate distillations are given : • 

Specific gravity 
at 15°. 
Initial boiling point 91°. 

Gasoline fraction 91° to 150° 39 per cent 0.770 

Kerosene fraction 150° to 300° 44.5 per cent 0.860 

Residium above 300° by difference 16.5 per cent. 

UNSATURATED HYDROCARBONS. 

The unsaturated hydrocarbons present in the crude oil and the gasoline 
and kerosene fractions were determined as follows: 

Twenty-five cubic centimeters of oil measured at room temperature 
were transferred to a 100 cubic centimeter, graduated glass-stoppered cylin- 
der, 25 cubic centimeters of sulphuric acid (specific gravity 1.'84) were 
added and the mixture shaken until no more reduction in the oil volume 
was observed. The volume of oil absorbed by the sulphuric acid was 
taken to represent unsaturated hydrocarbons. Under the conditions of 
the test the crude oil, the gasoline and kerosene fractions contained 30, 
16 and 24 per cent of unsaturated hydrocarbons, respectively. 

NATURE OF THE UNSATURATED HYDROCARBONS. 

One liter of the crude oil was shaken for eight hours with 0.5 liter 
of a mixture consisting of 1.5 volumes of sulphuric acid (^ecific gravity 
1.84), and 1 volume of fuming sulphuric acid (50 per cent SO.,). Under 
these conditions 700 cubic centimeters of oil were separated, leaving 300 
cubic centimeters, or 30 per cent of the original volume, absorbed by the 
acid. 



6 RICHMOND. 

The acid layer, for the separation of the aromatic from the other 
unsaturated hydrocarbons, was largely diluted with water, calcium car- 
bonate added to slight excess, and the brown solution of calcium sul- 
phonates separated from the precipitated calcium sulphate by filtration 
and evaporated to dryness on a steam bath. 

An attempt to recover the benzene homologues by steam distillation 
of the sulphonic acid salts in the presence of phosphoric acid alone ^ was 
unsuccef^sful. 

Approximately 500 grams of the dry calcium sulphonates were mixed with 500 
grams of glacial phosphoric acid, (specific gravity 1.715) and subjected to ex- 
haustive steam distillation, but only about 5 cubic centimeters of oil were 
recovered from a liter of aqueous distillate. The maximum temperature attainable, 
namely, the boiling point of glacial phosphoric acid (130°), was insufficient to 
cause hydrolysis of the sulphonic acids present. 

At this stage, about 200 grams of concentrated sulphuric acid were added to 

the contents of the flask and the distillation continued. The temperature rapidly 

rose to 140° and hydrocarbons began to distill freely. The temperature gradually 

rose to 170°', at which point the distillation was practically finished. A total 

volume of 148 cubic centimeters of oil, approximately 50 per cent of the unsaturated 

hydrocarbons present in the crude oil, was obtained. The oil was washed with 

water dried over calcium chloride and separated by distillation into the following 

fractions : 

Per cent by 
Initial boiling point, 130° volume. 

First fraction, 130° to 150° 25 

Second fraction, 150° to 200° 38.5 

Third fraction, 200° to 250° ' 17 

Fourth fraction, 250° to 300° * 17 

Residue by difference 7.5 



Total 100.00 

Each of the above fractions was nitrated and the derivatives obtained showed 
a close homologous series of aromatic hydrocarbons beginning with xylene, CgHio. 
Neither benzene nor tuolene nor any of the naphthalene series were found. 

THE REFINED OIL. 

The oil remaining from the acid treatment^ consisting of 70 per cent 
of the original crude petroleum by volume, was further refined by 
thorough agitation with strong caustic soda, and then washed with water 
and dried over calcium chloride. The specific gravity of the dry oil was 
0.793 at 15° C, at which temperature it became somewhat cloudy, due 
to the separation of solid hydrocarbons. At 0° C. it was quite viscous and 
at — 5° it was completely solid. 

*Friedel and Crafts. Compt. rend. Acad. set. (1889) 109, 95. 



TAYABAS PETROLEUM. 7 

One hundred cubic centimeters of the refined oil measured at 15° were 
distilled by Engler^s method with the following results : 

Per cent by volume, Specific gravity 
Initial boiling point, 80° cubic centimeters. at 16*. 

First fraction, 80° to 150° 40 0.746 

Second fraction, 150° to 300° 45 0.805 

Third fraction, 300° to 350° 7 

Residue above 350° by difference 8 



Total 100 

Paraffine W3is determined by Holde's method ° on the residue remaining from 
the distillation of the crude oil to the temperature of 260°, it being shown 
previously that no solid hydrocarbons were carried over at this temperature. 
The residue from 100 cubic centimeters of oil was transferred to an Erlenmeyer 
flask, diluted v^^ith an equal volume of ether and absolute alcohol (1 to 1 by 
volume), the solution cooled to — 20° and filtered with suction through a double 
filter also cooled to — 20°. The precipitated paraffine was washed with cold ether- 
alcohol mixture until free from oil and almost pure white in color. It was then 
dissolved from the filter with hot benzene and evaporated to constant weight at 
100°. The amount found was 6.775 grams or 8.1 per cent, calculated upon the 
weight of 100 cubic centimeters of the oil at 15°. 

Asphalt was determined by Holders method,' which consists in shaking 1 cubic 
centimeter of residue with 40 cubic centimeters of light petroleum; after 
forty-eight hours standing, the precipitated asphalt is filtered and weighed. 
The gasoline for this determination was Kahlbaum's ligroin, purified with 
sulphuric acid and distilled over sodium, using the fraction boiling between 65° 
and 95°. The amount of asphalt found was 0.08 per cent. 

SUMIMARY. 

1. Tayabas oil is characterized by its comparatively high proportion 
of volatile hydrocarbons, by having a paraffine base and by being sulphur 
free. In these respects it resembles Pennsylvania petroleum. 

2. It contains a greater proportion of volatile constituents boiling 
under 150° than most crude petroleums, and herein lies its greatest local 
value. 

3. It contains 30 per cent of unsaturated hydrocarbons, which are 
removed by the acid refining process, again resembing Pennsylvania pe- 
troleum, which, according to Engler,^ contains 28 per cent by volume 
soluble in sulphuric acid. 

4. It is essentially a paraffine petroleum, which, according to Peck- 
man,^ is the best kind of oil for fuel, illumination or lubrication. 

5. The crude petroleum examined was fresh seepage oil pumped from 
a 40-meter well, and is therefore considered representative of the oil 
from the oil-bearing strata of its source. 

*Untersuch. d. Mineralole u. Fette. Berlin, 2nd ed. (1905), 21. 

* Log cit. 

^ Ber d. deutch. Chem Qes., Berlin (1895), 28, 2501. 

*Journ. 8oc. Ohem. Ind. London (1900), 19, 1001. 



THE COMPOUNDS WHICH CAUSE THE RED COLORATION 
OF ANILINE: I. THE EFFECT OF OXYGEN AND OZONE 
AND THE INFLUENCE OF LIGHT IN THE PRES- 
ENCE OF OXYGEN. 



By H. D. GiBBS. 

{From the Laboratory for the Investigation of Foods and Drugs, Bureau of 

Science, Manila, P, I.) 



The cause of the formation of the red color in aniline has been attrib- 
uted to impurities and to oxidation. The literature does not contain 
as great a mass of confused and inacurate statements upon this subject 
as upon the question of the coloration of phenol.^ 

A. Rosenstiehl * remarks that, as is well known, aniline is turned brown in 
the air. On neutralizing with acid the color turns to rose due to the formation 
of pseudorosaniline by oxidation. 

He exposed aniline in a balloon flask to the action of air in the sunlight for 
three months and distilled the contents of the flask in vacuo. The distillate was 
pure aniline and the residue in the flask on acidification was colored rose, due to 
rosaniline. 

A. Bidet ^ states that pure aniline and its homologues are colorless when 
freshly distilled, become yellow after, some days, but never acquire the dark tint 
of the impure compounds. He ascribes the color to thiophene and thiophene 
derivates and believes that the coloration of benzene compounds is a much more 
delicate test for thiophene than the isatin reaction. 

P. Werner * also recalls the well-known fact that when aniline is exposed to 
air and light it becomes colored, first yellow, then red and finally brown. He 
placed samples of pure, colorless aniline in vacuo, in oxygen and in air in the 
diffused daylight for three and one-half months and corresponding samples in 
the dark. Those in vacuo remained colorless. The samples in oxygen and air 
in the daylight were colored and absorbed considerable quantities of oxygen, 
while in the dark the coloration was slight and the oxygen absorption in the 
tube containing oxygen was "very slight," and in the tube containing air 
*• inappreciable." 

From these data he concludes that the coloration is proved to be due to 
oxidation caused by the combined action of light and oxygen (air). 

^Gibbs: This Journal, Sec. A (1908), 3, 361. 
""Compt. rend. Acad. sci. (1876), 82, 380. 
""Gompt. rend. Acad. sci. (1889), 108, 620. 
*Journ. Soc. Chem. Ind. (1890), 9, 278. 



10 GIBBS. 

From his experiments he seems hardly to be justified in stating that 
the combined action of light and air are necessary^ for, while the aniline 
in vacuo remained colorless both in the dark and in the light, in the 
presence of air or oxygen a coloration was noticeable in the dark as well 
as in the light; so that the action of light, in the presence of oxygen, 
would simply seem to accelerate the reaction. 

Hantzsch and Freese ^ believe that the color is caused by a thiophene derivative. 
They state that aniline to which 0.5 per cent of thiophene has been added 
distills with an intensely yellow color, while a sulphur-free aniline distills without 
color and remains colorless for weeks. 

Since the literature offers only two explanations, impurities and 
oxidation, for the cause of the coloration, the various text-books have 
chosen either one or the other. The following are a few of the statements 
found. 

Meyer and Jacobson ^ state that the usual preparations of aniline are quickly 
turned yellow to brown by light and air. The color appears in the presence of, 
and depends upon, small amounts of a sulphur compound. 

Roscoe and Schorlemmor^ state that aniline "on exposure to light and air 
becomes brown and the more impure it is the more rapidiy does this coloration 
take place." 

Allen ^ states: "Aniline becomes yellow or brown on exposure to air and light, 
especially at elevated temperatures, a resinous body being ultimately formed. 
The change is due to oxidation and does not occur in vacuo or in the dark." 

Holleman-Walker ® state: "Aniline is a colorless liquid, and, unless perfectly 
pure, turns brown in the air, the color change being probably due to the presence 
of traces of sulphur compounds." 

I have found that the cause of the coloration of pure aniline in the 
presence of oxygen is oxidation. Azobenzene, dianilinoquinone, dianil- 
inoquinoneanil, and azophenine have been isolated from the red aniline, 
but are by no means the only substances produced by the oxidation. Since 
the oxidation, because of the presence of the quinone derivatives, has been 
shown to go as far as quinone, it is reasonable to expect that intermediate 
products are formed at least as transition compounds, and that these exist 
in the aniline solution if they escape further oxidation and do not readily 
form condensation products. 

The fact that the three quinone condensation products, dianilinoqui- 
none, dianilinoquinoneanil, and azophenine, have been isolated would 
seem to indicate that they all exist in the aniline solution. This, how- 
ever, is by no means proved for the reason that the methods employed for 
their isolation may convert, at least in part, one derivative into another. 

^Ber. d, deutschen chem. Oes. (1894), 27, 2529 and 2966. 

•Organische Chemie, Leipzig (1902), 2, 168. 

^Treatise on Chemistry, New York (1897), 3, pt. 3,199. 

* Commercial Organic Analysis, Philadelphia (1892), 3, pt. 2, 43. 

''Text-book of Organic Chemistry, New York (1906), 370. 



COMPOUNDS CAUSING RED COLORATION OF ANILINE: I. 11 

EXPERIMENTAL. 
ANALYSIS OF THE ANILINE COLORED BY THE PRESENCE OF OXYGEN AND SUNLIGHT. 

Samples of 100 cubic centimeters of aniline in 200 cubic centimeter bottles 
were placed in the direct sunlight and constantly agitated for about one month. 
The stoppers were removed from time to time and the air over the liquid changed. 
The aniline quickly became light yellow and finally, after passing through the 
intermediate stages, an intense deep red. 

Separation of the reaction products was affected in several different ways. 
On pouring* 20 cubic centimeters into a large volume of very dilute sulphuric 
acid, the aniline all dissolves as aniline sulphate. A residue which does not 
dissolve imparts to the mixture a reddish-brown color. After filtration the almost 
colorless solution changes to yellow and finally assumes a red color on standing. 
From the insoluble portion on the filter there was isolated 2, 5-dianilinoquinone. 

Other separations were made by extracting the dilute sulphuric acid solution 
of the colored aniline, without removing the suspended precipitate, repeatedly 
with small quantities of ether. The larger proportion of the precipitate went 
into the ether, forming a red solution. The etherial solution was filtered and 
evaporated in a vacuum dessicator. The residue separates in layers, one red 
and another a dark purple, almost black. The red portion was soluble in alcohol 
and was separated by this solvent. Dianilinoquinone was identified in the 
insoluble portion. The red alcohol solution lost its color on treatment with 
nascent hydrogen and other reducing agents. On stow evaporation yellow-brown 
needles of dianilinoquinoneanil separated. Another portion of the alcohol solution 
was warmed until the alcohol had evaporated. The red residue was distilled 
with steam. A few drops of an orange red oil, which solidified to orange-red 
crystals, passed over with the steam. These crystals were readily reduced in 
alcoholic solution to hydrazobenzene. Azobenzene must, therefore, be present. 

Other separations were made by slowly pouring 20 cubic centimeters of the 
aniline into 100 cubic centimeters of 50 per cent acetic acid. On cooling the solu- 
tion in ice, a large number of small crystals separated. When viewed with the 
microscope they were recognized on sight as azophenine. The marked tendency to 
twinning is quite characteristic. A very few crystals of dianilinoquinone were 
also observed. 

Twenty cubic centimeters of the colored aniline were distilled in steam until 
the volatile compounds had passed over. The red residue in the flask was 
filtered and the precipitate dissolved in a small quantity of hot alcohol. On 
standing, crystals of azophenine separated. 

2, 5-dianilinoquinone was identified by its crystalline form, high melt- 
ing point and coloration produced with concentrated sulphuric acid. 
Since the descriptions of this compound given in the literature are in some 
points conflicting, a study of some of its properties has been made. 

The appearance of the crystals is described by A. W. Hofmann,^® who first 
prepared the compound, as "reddish-brown almost metal lustrous scales;" by 
Knapp and Schultz ^^ as "brown-red, metallic glistening, leaflets;" by Zincke 
and Hebebrand^^ as "small, bluish-violet leaflets," and by Nietzki and Schmidt" 
as "steel-blue, glistening leaflets." 

^''Proc. Roy, 8oc. (1863), 13, 4. 
^^Ann. d, Ghem, (Liebig), (1881), 210, 179. 
^^Ber. d. deutschen chem. Ges. (1883), 16, 1556. 
'^hid. (1889), 22, 1656. 



12 



GIBBS, 



These observations are more or less correct, depending upon the light 
by which the crystals are viewed. By reflected light the appearance is 
usually a brilliant, blue-violet, while by transmitted light the predominat- 
ing color is yellow to reddish-brown. Even these shades vary greatly 
under the microscope. From the constitution of the compound, as 
developed by Nietzki and Schmidt, one naturally suspects that its color 
will be more apt to be of the quinone order and that the description 
yellow-brown to red-brown is most appropriate. Some difficulty was 
experienced in obtaining the crystals in their true form, probably due 
to slight impurities. By washing many times with boiling alcohol, 
cold alcohol, boiling ether and cold ether consecutively, after recrystalli- 
zation from aniline, quite pure crystals were obtained. I am indebted 
to Dr. W. D. Smith, chief of the division of mines, of this Bureau, for 
the following description of the crystals : 

The compound crystallizes in the tetragonal system and occurs in very thin 
plates, generally under a millimeter in length. The usual forms are prism, 
pyramid and base, diamond and square. In- mass and by reflected light these plates 
have a decided purple hue. By ordinary transmitted light many of them are 
quite colorless. The crystals are trichroic. The accompanying fig. 1 gives the 
form, and figs. 2 and 3 show the pleochroism : 



89030/ 



Fig. 2. 



Fig. 3. 



These plates show parallel extinction in polarized light. The index of re- 
fraction is medium and the double refraction is low. First order colors. 

The melting point of the compound is stated by Nietzki and Schmidt 
to be above the thermometer scale, while Zincke and Hebebrand say that 
it does not melt and sublimes without decomposition. I have found 
that the crystals melt readily and recrystallize in broken and imperfect 
forms on cooling. Probably slight decomposition takes place. The melt- 
ing or softening begins at 338° and the purest samples obtained were 
completely melted at 342°, uncorrected. A nitrogen determination of this 
subst-ance gave 9.76 per cent, while the theoretical value is 9.65 per cent. 
In concentrated sulphuric acid the coloration is stated by Knapp and 
Schultz to be ^^fuchsin red,^^ and by Nietzki and Schmidt ^^violet.^' I 
have observed that the former statement more accurately describes the 
color. 



COMPOUNDS CAUSING RED COLORATION OP ANILINE: I. 



13 



2, 5-dtaniUnoquinonediaml (azophenine) was identifid by its very 
characteristic crystalline form, and the coloration produced with con- 
centrated sulphuric acid. The crystals employed for the purpose of 
comparison were made by oxidizing aniline with iodic acid ^* in the pres- 
ence of the required concentration of acetic acid. The compound was 
recrystallized from aniline. 

0. Lehmann^" describes the crystals as monoclinic needles from aniline which 
unite extraordinarily often to twins which cross, with pronounced tendency to 
bending, and which are very dichroic. Beilstein^' gives the description, garnet- 
red leaflets. 

I am indebted to Dr. W. D. Smith for the following description of 
the crystals. The accompanying figures show the characteristic forms. 




Fig. 4. 



[xlinction 

Fig. 5. 




Fig. 6. 



A' 



V 

Fig 



The compound crystallizes in the monoclinic system and by reflected light 
appears garnet-red in color. The crystals, which appear almost invariably as 
shown in fig. 4, measure much under a millimeter in length, one of the longest 
in a group measuring 0.18 millimeter. They are very pleochroic, varying from 
colorless through yellow to garnet-red. The crystals usually occur in halves, 
the normal crystallization being as shown in fig. 7. In fig. 6, a twin is shown, 
the twinning angle being 31°, the same as the extinction angle. The index of 
refraction is medium. The double refraction is strong, the interference colors 
are reds and blues of the second order. 

All of the compounds which have been isolated are soluble in aniline 
with the production of colored solutions. The following is a brief 
study of the solubilities : 

Azophenine is somewhat soluble in alcohol and aniline and is readily crys- 
tallized from either solvent. The solutions are brilliant red. 

2, 5-dianilinoquinoneanil is soluble in alcohol and aniline, the solutions being 
brilliant red. 

2, r5-dianilinoquinone is practically insoluble in alcohol and soluble in aniline, 
forming a brilliant red solution. 

Azobenzene is very soluble in aniline forming a reddish-yellow solution. 

The power of the first three compounds of coloring colorless aniline is intense 
and exceeds by many fold the coloring power of the latter compound. 

^* Ostrogoyich and Silbermann: Chem. Ahs. (1908), 2, 1433. 
^Uahresh. d. Chem. (1882), 369. 
'*Handbuch, Org. Chem. (1897), 3, 341. 



14 GIBBS. 

With concentrated sulphuric acid azophenine gives a reddish-purple resembling 
the initial color obtained on dropping a crystal of morphine into the formal- 
dehyde-sulphuric acid reagent, 2, 5-dianilinoquinoneanil a salmon-red " and 2, 5- 
dianilinoquinone a fuchsin-red color. 

THE OXIDATION OF ^ANILINE. 

The oxidation of aniline with various oxidizing substances has been 
productive of many and various reaction products. The reagents which 
seem to have the most intimate bearing upon the reactions which ordi- 
narily produce the coloration of aniline are oxygen^ ozone and peroxides. 

Otto" passed ozone through a solution of aniline in benzene at 15° and 50°, 
and found the principal products to be quinone and azobenzene. 

Leeds " oxidized aniline acetate in acetic acid solution by means of hydrogen 
peroxide and obtained azobenzene. * 

Schunck and Marchlewski ^ and others oxidized aniline by means of hydrogen 
peroxide in acetic acid solution and found azobenzene, dianilinoquinoneanil and 
ammonia. They express the reaction 

4 CoH,NH,-f 4 H202=C24H,;[^30+7H20-f NH3. 

In the presence of stronger acids, indulin compounds are formed. 

Freer and Novy^^ oxidized aniline with benzoylacetyl peroxide in etherial 
solution and obtained azoxybenzene as the principal oxidation product. 

Istrati ^^ studied the products obtained by aspirating air through boiling 

O 



aniline. Compounds to which he assigns the formulas 0^(1 « ^i ^^^ 

VN— QH5/3 

O 
(C«H5-NH)3 : CfiHj, • • CgH., \ (NH-C^H,)., are described. 

It is extremely probable that quinone condensation products which he failed to 
identify were formed. 

EXPERIMENTAL. 

THE EFFECT OF SUNLIGHT UPON PURE, DRY ANILINE IN CONTACT WITH DRY 

ATMOSPHERIC AIR. 

Since impurities have been shown to produce a coloration of aniline., 
and moisture often influences to a great extent the chemical action of 
lights the effect of the sunlight upon pure, dry aniline enclosed with 
purified, dry atmospheric air was next studied. 

Pure aniline was obtained from purified and re-crystallized acetanilid 
by treating with potassium hydroxide ^ solution, distilling with steam, 
and extracting with ether. 

The etherial solution was dried over potassium hydroxide and fractioned 

"Zincke: Ber. d. deutschen chem. Ges.y 1885, 18, 787, says a blood-red color. 

^* Ann. chim. et phys. (1898), 13, 138. 

"J9er. d. deutschen chem. Gcs. (1881), 14, 1382. 

""Ihid. (1892), 25, 3574. 

''Am. Chem. Journ. (1902), 27, 178. 

"^Compt. rend. Acad. sex. (1902), 135,* 742. 



oompoCnds causing rei> coloration of aniline: I. 16 

several times. The aniline thus obtained had a constant boiling point. As a 
further precaution to avoid the presence of sulphur compounds, a portion of this 
product was further purified by the method of Hantzsch and Freese** by boiling 
with acetone. The final drying was accomplished by distilling with a small 
piece of metallic sodium in the manner previously described** and only the 
middle refraction employed. This was received in the flask / of the apparatus 
shown in fig. 1,^' a small piece of sodium introduced, the tube sealed and tho 
procedure employed in the investigation of phenol was carried out in the same 
manner. In some respects aniline is more easily handled than phenol. 

The liquid thus obtained showed no color visible to the eye. When 
oxygen was excluded by inclosing it with an atmosphere of an indifferent 
gas, it remained unchanged for two months, as long as the observation 
lasted, in the diffused light of the laboratory. 

When scaled in thin glass tubes with purified, dry atmospheric air, 
consisting only of indifferent gases and oxygen, the aniline colored ra- 
pidly in the sunlight, a few minutes^ exposure being sufficient to produce 
a light red color. The change takes place with remarkable rapidity. 
When the sun is near the zenith at noon in either the months of April, 
]\Iay or August I have watched a tube of pure, colorless aniline darken 
and assume a decided red shade in less than ten minutes. The tempera- 
ture is an important factor of the rate of the reaction. 

Since one of the first products to be formed is water, the reaction 
products in general will be the same as those previously described in the 
experiments where no effort was made to exclude) moisture. This experi- 
ment merely proves that moisture is. not necessary to start the reaction. 

THE ACTION OF OZONE ON ANILINE. 

On contact with ozonized oxygen, colorless dry aniline behaved in ap- 
pearance in the same manner as liquid phenol. Coloration was instantly 
produced and gradually deepened from yellow to a very dark red as the 
action progressed. The presence or absence of the ion filter, G, described 
in the work on phenol, produced no visible effect. In the beginning 
no carbon dioxide was envoi ved, but as the reaction proceeded large quan- 
tities were given off. The aniline finally became so saturated with the 
reaction products that crystals separated and eventually the whole became 
almost a solid mass of crystals. When viewed by the microscope these 
were seen to be dianilinoquinoneanil. No other crystals were detected. 

Aniline which had not been purified with any great degree of care 
behaved, in the presence of moisture and ozone, as far as was observed, 
in the same manner. Dianilinoquinoneanil was the principal product 
formed. 

^ hoc. cit. 

2* The Oxidation of Phenol; This Journ., Sec, A. (1909), 4, 137. 

''Ihid., p. 136. 



16 GIBBS. 

THE COLORATION OF ANILINE IN THE PRESENCE OF OXYGEN IN THE DARK. 

Tubes of pure, dry aniline in contact with purified dry atmospheric 
air, prepared with great care as previously described, were wrapped in 
several coatings of tin foil to exclude the light, and heated in a steam 
bath in the same manner as described in the work on phenoL^^ The 
coloration appeared in a considerably shorter space of time than that 
required for the coloration of phenol and became darker with time. At 
room temperature, 30°, the purest samples of aniline become colored 
on long standing in the dark. The first shade of color is yellow, which 
slowly changes on long standing to light red. 

. The next paper will deal with the coloration of aniline and some 
light reactions in the absence of oxygen. 

SUMMARY. 

1. A brief review of the literature of the coloration of aniline is 
given. 

2. While there can be no doubt that certain impurities will color 
aniline, the coloration of the pure compound, in the presence of oxygen, 
is due to oxydation. 

3. The coloration goes on in the dark and is accelerated directly with 
the temperature. 

4. Sunlight greatly accelerates the reactions involved. 

5. The presence of moisture or impurities are not necessary factors. 

6. Among the products of the light reactions there have been isolated, 
azobenzene, 2, 5-dianilinoquinone, dianilinoquinoneanil and azophenene. 

7. Ozone reacts instantly with pure, dry aniline or aniline and mois- 
ture, producing a brilliant red solution. The principal product of the 
reaction is dianilinoquioneanil. Carbon dioxide is copiously evolved. 

8. The experiments argue against any increased chemical activity of 
the oxygen due to the oxygen gas ions. 

9. The rate of the color formation in aniline is much greater than 
in phenol. 

10. It is evident that the purest aniline is prepared by distillation at 
reduced pressures and in atmospheres of indifferent gases. 

^""Loc. cit. p. 149. 



ERRATA. 



In the article on the "Oxidation of Phenor' in This Journal. Sec, A (1909), 
4, 134, footnote 11, 97 should be 91; page 150, lines 10 and 11 from the bottom 
of the page, the first shorter and longer in each line should be transposed. 
93401 2 17 



: :.J. ™::u„::w,-;,;2;i*;ftoS 



A LIQUID CONCAVE MIRROR. 



By Clinton H. Hulbtjbt^ and R. F. Bacon. 
{From the Chemical Laboratory ^ Bureau of Science,) 



The recent publications of Wood ^ on liquid mirrors induces us to 
describe a similar mirror with which one of us has been experimenting 
for several years.^ 

The general idea of using centrifugal force on liquids to form perfect 
paraboloid surfaces seems to be very old, but its practical application is 
of quite recent date. 

The principal problem which we undertook to solve was largely a 
mechanical one. To obtain a perfect mirror it is obviously necessary 
that the vessel containing the liquid should rotate at a uniform speed 
and without vibration. The apparatus which we finally evolved to meet 
these conditions is shown in Plate I. The mirror proper is formed in the 
innermost of three cylindrical copper vessels, 22, 30, and 36 centimeters 
in diameter, respectively. The two smaller vessels have double sides 
with an air space between the walls to give greater buoyancy. In most 
of our experiments the liquid used was a heavy, black paraffine oil, and 
each vessel was filled about one-quarter full with this oil. For most 
purposes the paraffine oil works quite as well as mercury. The two inner 
cylinders have at the center of the bottom, double cones of steel which 
causes these vessels to be held in the exact center of the apparatus by 
means of the magnetic solenoid A. To rotate the pans, an ordinary 
laboratory motor of one-eighth horsepower was used. This was equipped 
with suitable gears to vary the speed of rotation and the power was 
transmitted to the pans by means of a belt, the whole apparatus being 
mounted on a heavy wooden board. As the oil used is quite viscous, 
any vibrations due to mechajiical imperfections in the bearings of the 
largest pan are absorbed by the oil, and not communicated to the other 
two pans. Similarly, any variations in the speed of the motor are 

^Formerly in the Quarternmsters Department, United States Army. 

* Astro Physical Journal (1909), 29, 165. 

•Mr. Hulburt first began this work in 1902 and the writer (Bacon) became 
associated with him early in 1906. A note was published in the Scientific 
American, 1907, p. 103. 

19 



20 HULBUET AND BACON. 

approximately compensated by the time the innermost mirror float is 
reached, as the two inner vessels are run by a liquid friction drive. 

In practice, this apparatus has given very satisfactory results. The 
images formed by a paraffin oil mirror are quite sharp and the focal 
length and magnification are very easily regulated by changing the speed 
of rotation by means of the set of gear wheels shown in the figure. 
With such an exceedingly simple apparatus we were able to secure good 
views of the craters of the moon. 

Plate II shows the film of an incandescent light bulb as shown by a 
mirror of 30 centimeters diameter, the rotation being such that the focal 
length was about two meters and the magnification about four times. 
It would seem that there would be neither great mechanical difficulty 
nor great cost in constructing very large mirrors of this type, and as the 
surface is shaped as a result of molecular forces, it is mechanically 
perfect, the only possible disturbances being due to vibration or variation 
in the rotation, both of which factors are practically eliminated by the 
arrangement of the apparatus used by us. 



ILLUSTRATIONS. 



Plate I. Apparatus used in forming a liquid concave mirror. 
II. Photograph of film of an incandescent light bulb. 

21 



DISTILLED LIQUORS: WHISKY (RYE, BOURBON AND 
SCOTCH), BRANDY (COGNAC) AND GIN. 



By W. C. Holmes. 

{From the Laboratory for the Investigation of Foods and Drugs, 
Bureau of Science.) 



I. A Study of the methods of analysis and some suggested improvements. 
II. Analyses of distilled liquors entering the port of Manila. 

T. A STUDY OF THE METHODS OF ANALYSIS AND SOME SUGGESTED 
IMPROVEMENTS. 

A considerable number of analyses of distilled liquors have recently 
been carried out at this laboratory and in the course of this work several 
departures from the methods adopted by the • Association of Official 
Agricultural Chemists have been considered advisable. 

In the determination of volatile acidity, the method of adding 25 cubic 
centimeters of water to 200 cubic centimeters of ohe sample and slowly dis- 
tilling off 200 cubic centimeters of distillate, has been replaced by one in which 
100 cubic centimeters of the sample are distilled with the direct flame until 
the volume has diminished to about 30 cubic centimeters and then the distillation 
is continued with steam, together with the direct flame, the volume of liquid 
in the distilling flask being kept nearly constant, until 500 cubic centimeters 
of distillate are obtained. Aliquot portions of the distillate are taken for the 
determination of alcohol, volatile acids, esters, and aldehydes. This distillate 
is too dilute for the accurate color imetric determination of furfurol. In many 
cases, however, this is of too little importance to necessitate a separate distillation. 

This modification of the method gives higher and more accurate values 
for the volatile acids and is less affected by minor variations in the 
manipulation. The determination of alcohol in a distillate obtained in 
this manner is sufficiently accurate for the ordinary purposes of analysis, 
and if necessary, in special cases, may be supplemented by a determina- 
tion according to the official method. 

The colorimetric method for the determination of aldehydes has not 
proven as satisfactory as the following procedure based on the method 
of M. Eipper.^ i 

One hundred cubic centimeters of distillate are pipetted into an Erlenmeyer 

N 
flask, 25 cubic centimeters of r^ sodium hydrogen sulphite solution added, the 

"■Monatsh, f. chem. (1900), 21, 1079. 

23 



24 HOLMES. 

flask tightly corked, shaken, and allowed to stand for one-half hour. The 
uncombined bisulphite is then determined by titration with ^^ iodine solution 
and starch indicator. Blank determinations are made in the same manner. 

The calculation of the aldehyde as acetaldehyde is made according to the 
formula. 

CO.— iodine solution used up X 550.7 

Grams acetaldehyde per 100 liters proof spirits = 

per cent alcohol in sample. 

The Allen-Marquardt procedure for the determination of fusel oil 
as given in the methods of analysis of the Association of Official Agri- 
cultural Chemists ^ has been modified in several particulars. Preliminary 
experiments with known amounts of amyl alcohol indicated that the 
method gave only about 80 per cent of the amyl alcohol present; a 
degree of efficiency corresponding very nearly to that reported by C. A. 
Crampton.^ This discrepancy is due in a large measure to the erroneous 
assumption that inorganic acids are present in the- distillate obtained 
after oxidation, which must be first neutralized by titration, employing 
methyl orange as indicator. 

Mann and Stacy* have shown that no inorganic acid is present in 
this distillate and that the acidity indicated by methyl orange is organic. 
In our method, the final distillation is made with steam, care being 
taken not to reduce to any great extent the volume of liquid in the' 
distilling flask. About 300 cubic centimeters of distillate are collected, 
titrated with phenolphthalein and the entire acidity considered as valeric 
acid. 

About twenty of these distillates were carefully tested for inorganic acidity 
with negative results. No hydrochloric acid was found even when the final washing 
of the carbon tetrachloride extract with saturated sodium sulphate was omitted. 
It seems probable that there is some loss of valeric acid during the oxidation, 
although the evidence in the literature upon this question is contradictory.'^ On 
the other hand, Mann and Stacy demonstrated that with temperatures above 
15.° 6 the separation of ethyl alcohol and amyl alcohol is incomplete and that 
an appreciable amount of ethyl alcohol is left in the carbon tetrachloride, and a 
corresponding amount of acetic acid found in the final distillate. Since the 
temperature of this laboratory varies very little from 30° it is probable that the 
positive error thus introduced more than compensates for any loss of acid. 

Mann and Stacy ° have devised a further modification of the method by substitu- 
tuting for the eight hours' oxidation with reflux condensers, on oxidation of four 
hours or even less in mechanically shaken pressure bottles, thereby eflfecting an 
economy in time and eliminating the possibility of loss of valeric acid. 

* U. 8. Dept Agric. Bur. Chem. Bui. 107 (1908), 98. 

*Pr6ced. 23d Ann. Convention of Association Oflf. Agr. Chem. Bui. 105, 
Bureau of Chem., Ul S. Dept. Agr. (1907), 20. 
*Joum. Soc. Chem. Ind, (1906), 25, 1125. 
' Schidrowitz, Analyst (1901), 31, 181; Loomis, loc. cit. Bull. 105. 22. 

* Loc. cit. 



DISTILLED LIQUORS. 25 

A more recent modification of the method^ as yet unpublished is dependent 
upon the use of a standard oxidizing solution and the determination of its 
loss in strength. 

Bedford and Jenks ^ have shown that while the method of separation of the 
alcohols in the Allen-Marquardt method is adequate .in the case of amyl alcohol 
it fails to give satisfactory results with the intermediate alcohols." 

They found that saturated calcium chloride solution gave excellent results with 
the two butyl alcohols and normal propyl alcohol, when substituted for sodium 
chloride solution in the washing. Very good results were obtained with all the 
higher alcohols which they tested, with the exception of isopropyl alcohol, a 
part of which only was extracted. They further state that practically the most 
serious disadvantage of the oxidation method is the length of time (three days) 
required for a determination. 

In the Bedford and Jenks method, which can be carried out in a day, the 
separated higher alcohols are converted into nitrites which are then decomposed by 
acid in the presence of potassium iodide and the liberated iodine titrated. They 
point out that their ester-iodine method gives the total higher alcohols; while the 
Allen-Marquardt method gives practically only the amyl alcohol, and since, as 
they showed experimentally, the ratio of amyl alcohol to the total higher alcohols 
differed in a marked degree in spirits subjected to different modes of distillation, it 
followed that by the supplementary employment of both methods, considerable 
light could, in many cases, be thrown on the nature of any sample under investiga- 
tion. For example, a genuine pot still whisky could easily be distinguished from 
a whisky made from rectified spirits. 

The interpretation of analytical results in the present work was based 
almost entirely on the data given by C. A. Crampton and L. M. Tolman 
in their paper on the changes taking place in whisky stored in wood.^^ 

This admirable and thorough research has made it possible for the 
analyst to pass judgment on the nature of American whiskies and sufficient 
data were obtained with whisky stored in uncharred packages to warrant 
the application of the same principles to Scotch whiskies. 

In the case of brandy less evidence was available, but from the analyses 
made by Mann and Kerton^^ and from those of the best samples at 
hand, it was possible to form some opinion of the composition of a genuine 
brandy. 

One of the tests which, according to A. McGill ^^ will serve to dif- 
ferentiate in many cases, between a genuine pot-still distillate and one of 
which the basis is rectified spirits, consists in the dilution of the distillate 

^ This method is being experimented with in some of the laboratories of the 
Bureau of Chemistry, U. S. Dept. Agr. 

"Journ. Soc. Chem. Ind. (1907), 26, 123. 

• They say "* * * although amyl alcohol by itself can be determined with 
a high degree of accuracy, the intermediate alcohols (normal butyl, isobutyl, and 
normal propyl ) are also returned, but with notably low results, whilst, in addition, 
isopropyl alcohol is theoretically undeterminable by the method, owing to its 
oxidation to acetone instead of an acid." 

^^Journ. Am. Chem. Soc. (1908), 30. 98. 

^-"Joum. Soc. Chem. Ind. (1907), 26, 450. 

^""Lah. of Int. Rev. Dept. Canada, Bui. No. 27. 



26 



HOLMES. 



to twice the volume of the original sample. The presence of volatile oils, 
in general characteristic of pot-still spirits, is then indicated by an opales- 
cence. In the latter part of the present work the appearance of the 
distillates made to determine volatile acids was noted and a correlation 
attempted of these indications with the nature of the spirits as indicated 
by analysis. At this greater degree of dilution an opalescence was ob- 
tained with nearly all samples, but this was apparently more marked in 
the case of spirits of pot-still origin, particularly in cases in which the 
product had received but little aging. 

II. ANALYSES OF DISTILLED LIQUORS ENTERING THE PORT OF MANILA AND 
SAMPLES OF THE SAME FOUND ON THE LOCAL MARKETS. 

Rye whisky. — Thirty- three anah^ses of rye whiskies were made of 
samples representing twenty-four brands. A general summary of results 
is given below, together with the average data obtained by Crampton and 
Tolman, for rye whiskies of four years^ aging. 

Table I. — Analyses of 2Jf rye whiskies. 



Maximum 

Minimum 

Average 

C. & T.'s average - 



Den- 
sity. 



.953 
.937 



Per- 
cent al- 
cohol. 



51.9 
38.1 



Total 
solids. 



675 
92 



Total 
acids. 



60.4 
16.2 



Vola- 
tile 
acids. 



51.4 

n.i 



Esters. 



Fusel Aide- 



oil. 



72.4 I 162.0 ! 
9.2 24.0 ! 



hydes. 

13.2 
1.2 



407 

185 I 



36.6 I 
65.9 i 



36.0 I 
69.3 ! 



59.5 
125.1 



5.7 
13.9 



All data except those for density and percentage of alcohol are given 
in terms of grams per hundred liters of hundred proof spirits. 

All except four of these twenty-four brands were colored with caramel, the 
detection of which was based on the Marsh method," and sugar or saccharine 
compounds had been employed in fifteen brands. Three brands which had been 
bottled in bond proved to be genuine pot-still whiskies of considerable aging and 
free from added coloring or flavoring matter and a fourth was a '^new" pot-still 
distillate to which coloring matter had been added. * The remaining samples were 
characterized by a low content of secondary distillation products. 

Bourbon whisky. — Seven samples of bourbon whisky were analyzed. 

Tabu: II. — Analyses of 7 bourbon whiskies. 





Den- 
sity. 

.949 


Per 
cent al- 
cohol. 


Total 
solids. 


Total 
acids. 


Volat- 
ile 
acids. 


Esters. 


Fusel 
oil. 

115.2 
23.8 


Maximum _^_ 


48.5 
44.0 


505 
227 


35.1 

18.6 


31.7 44.5 
13.7 9.2 


Minimum 


.943 


Average _ 


.946 


46.5 


391 
151.9 


27.1 
58.4 


, 19.3 



32.3 
53.5 


58.1 
123.9 


C. & T.'s average 









^^Journ. Am. Chem. 8oc. (1908), 30, 100. 



DISTILLED LIQUORS. 



27 



All of these samples were colored with caramel with the exception of one, which 
had practically no color, and all except one contained sugar or sugars. One sample 
only was a genuine pot-still whisky, although not well aged. 

Scotch whisky. — Forty-eight analyses of Scotch whiskies were made x>i thirty- 
seven different brands. Since no reliable information concerning the composition 
of Scotch whiskies was available the analysis of one of the best samples tested 
has, for purposes of comparison, been included in the following table. 

Table III. — Analyses of 37 Scotch whiskies. 



Maximum 

Minimum 

Average 
Best sample — 



Den- 
sity. 


F*er 
cent al- 
cohol. 


Total 
solids. 


Total 
acids. 


Volat- 
ile 
acids. 


Esters. 


Fusel 
oil. 


Alde- 
hydes. 


.946 
.930 

.939 
.938 


53.7 

44.8 

49.1 
49.3 


452 

56 

159 

168 


60.8 
4.7 


58.9 
2.5 


42.4 
5.9 


183.0 
29.0 


16.3 
3.0 


23.7 
60.8 


15.7 
53.9 


21.1 
42.4 


69.6 
96.6 


8.5 
12.6 



All of these samples were colored with caramel, although in duplicate samples 
of two of the brands none was found. The amount of color was slight and its 
presence is attributed by the distillers to the practice of aging Scotch whiskies 
in sherry casks. Only two samples contained sugars. Of all, only two appeared 
to be genuine, well-matured pot-still whiskies; in two cases the evidence was 
sufficiently doubtful to prevent their classification as young whiskies, and five 
brands were clearly new whiskies. The amount of secondary products of distilla- 
tion present in the remaining brands was small. 



Twelve samples of brandy were analyzed of which ten were labeled Cognac. 
The analysis of one of the best brandies received, which will be found to be 
in substantial agreement with the data given by Mann and Kerton** has, for 
purposes of comparison been included in the following table. 



Table IV. — Analysis of 12 brandies. 





Den- 
sity. 


Per 
cent al- 
cohol. 


Total 
solids. 


Total 
acids. 


Vola- 
tile 
acids. 


Esters. 

96.1 
20.5 


Fusel 
oil. 


Alde- 
hydes. 


Maximum _ 


.949 
.931 


53.7 
47.3 


2,189 
235 


83.6 
17.4 


63.8 
6.2 


150.6 
12.4 


15.4 
3.0 


Minimum 


Average 


.942 
.943 


49.7 

48.5 


1,170 


46.6 


31.0 

25.8 


63.5 
75.3 


71.1 
150.6 


8.0 
7.1 


Good brandy 


709. 37. 





All these samples were colored with caramel, and all but one contained sugars. 
The sample from which the comparative analysis was obtained was considered a 
genuine, but rather immature brandy. 

Oin. — Thirteen brands of gin have been analyzed, of which four were of Holland, 
five of English, and four of native origin. The average data obtained from these 



^* Loc. cit. 



28 



HOLMES. 



analyses are included in the following table together with the average figures 
of several analyses compiled by Kohig.^^ 

Table V. — Analyses of 13 gins. 



Origin of gins. 


Den- 
sity. 


Per 
cent 
alco- 
hol. 

46.6 
32.0 
51.1 

51.5 
46.8 


Total 
solids. 

43.5 

118.3 

2.6 

271.2 
27.1 


Total 
acids. 


Vol- 
atile 
acids. 


Esters. 


Higher 
alco- 
hols. 


Alde- 
hydes. 


Furfu- 
rol. 


McGill's opa- 
lescence test. 


Holland _ 

Native 


.945 
.963 
.938 

.950 


4.7 
8.4 
1.1 

1.9 
50.1 


3.9 
4.5 
1.1 

1.7 


21.3 
78.9 
8.9 

1.1 
119.2 


47.1 
57.4 
33.0 

26.5 
181.5 


2.5 
6.4 
2.1 

1.7 
11.0 


1.0 

None. 

None. 
0.9 


Very faint. 

Very faint. 

None. 

None. 


English 


English Long 
Tom 

Konig's figures _ 









It seems probable that the samples received at this laboratory were manufac- 
tured from rectified spirits and flavored with oil of juniper or some similar 
compound. In the absence, however, of any established standards for gin, and 
indeed of any accurate knowledge of the chemical composition of that liquor, 
very definite conclusions are unwarranted. 

In conclusion it may be stated that out of some eighty brands of 
whiskies and brandies not more than ten were found to be genuine, 
matured, pot-still liquors; and of these ten, five were probably not strictly 
mature, one contained sugars and seven were colored in the sense of 
containing caramel. The term "pot-still liquor^' has, throughout this 
paper, been employed in discription of liquors distilled in such a manner 
that the distillate contains a considerable amount of the secondary dis- 
tillation products congeneric with ethyl alcohol, which give to the liquor 
what may properly be termed its natural flavor. The .term "straight 
whisky" is now in general use in qualifying a whisky of this nature, which 
has been aged at least four years in wood, without the addition of caramel 
or sugars, and of such only three brands were encountered. The majority 
of the samples analyzed were taken from the retail trade, although a 
number were from shipments at the custom-house, and evidence was 
obtained showing that in certain cases the character of the liquor had 
been changed by the addition of rectified spirits after its importation 
into the Philippine Islands. 

"Konig, Chemie der Menschlichen Nahrungs-^und Genussmittel, I, 1408. 



ON THE DETERMINATION OF ALDEHYDES IN 
DISTILLED LIQUORS.' 



By Mariano Vivencio del Rosario. 
{From the Chemical Laboratory ^ Philippine Medical Schooly Manila, P. I.) 



The colorimetric method for the determination of aldehydes in dis- 
tilled liquors, provisionally adopted by the Association of Official Agri- 
cultural Chemists,^ is based upon the action of aldehydes toward Schiff's ' 
reagent. This method is identical with the one given by Tollman and 
Trescott * which, in its turn, is a modification of the method extensively 
used by Girard and Dupre.'^ 

Two volumetric methods for the determination of aldehydes have 
been proposed. Seyewetz and Gibello showed the possibility of titrating 
formaldehyde and its polymers with a solution of normal sodium sulphite 
made neutral to phenolphthalein by a few drops of acid. The reaction 
depends upon the fact that normal sodium sulphite, in tbe presence of 
acid, is converted into sodium hydrogen sulphite which forms an addition 
product with the formaldehyde. As long as the whole amount of formal- 
dehyde has not been changed to the bisulphite combination, the solution 
remains neutral in spite of the acid added. Seyewetz and Bardin ® have 
attempted to apply this method to aldehydes in general. 

The volumetric method of Eipper ^ is similar in principle and applica- 
tion. The amount of aldehyde is calculated from the amount of po- 
tassium bisulphite required to form the addition product as shown by 
titration with standard iodine solution. 

^ Abstract of the thesis presented to the Ogden Graduate School, University 
of Chicago, in candidacy for the degree of S. M. 

^ Bureau of Chemistry, U. S. Dept. Agric, 1 07, 96. 

^Compt. rend. Acad. sci. (1865), 61, 45. Ibid., (1887), 105, 1182. Centralhl 
f. Chem. (1905), 1305. Moniteur Scientifique (1890), 893. Ch. Girard et 
A.' Dupr^. Analyse des Matieres Alimentaires. Paris: V^* Ch. Dunod & P. Vice. 
(1894), 278. Ch. Girard et L. Cuniasse. Manuel Practique de 1' Analyse des Al- 
cohols et des SpiHtueux. Paris: Masson et C'*". (1899), 172. 

Wourn. Am. Chem. 8oc. (1906), 28, 1624. 

'Analysis of Potable Spirits, p. 30. 

^Bull Soc. Chim. de Paris. (1905), 33, 1000. 

' Monatsh. f. Chem. (1901), 21, 288. 

20 



30 VIVENCIO DEL ROSARIO. 

The fundamental equation is as follows : 

CH^' CHO + KHSOa-» CH^- CH<^^j^ 

Since considerable difficulty has been experienced in this laboratory 
in obtaining accurate results by the official method, probably due in a 
large measure to the temperature factor, the method of Eipper has been 
employed to a considerable extent. The following experimental data 
show that the method gives accurate results. 

EXPERIMENTAL. 

Preparation of the solutions. — Since practically all the aldehyde deter- 
minations which the commercial analytical chemist has to deal with are 
in alcoholic beverages, the experiments have been carried out on solutions 
of aldehyde in 50 per cent alcohol. 

Pure alcohol. — Although Yassey ^ says that the purest commercial 
cologne-spirits is sufficiently pure for use, it always contains some impu- 
rities. Of these the high-boiling, part, i. e., fusel oil, can be removed 
with considerable ease. The method of Winkler ^ as modified by Dunlap, 
using silver oxide, failed to give an alcohol free from aldehydes. Each 
100 cubic centimeters of distillate was tested for aldehydes with rosanil- 
ine-bisulphite solution and the characteristic reaction for aldehydes ob- 
tained with each portion tested. 

A second method of Girard and Dupre ^^ for the purification of alcohol 
is that recommended by Tollman in his article on whisky.^ ^ It failed 
to give satisfactory results in my hands until W. L. Lewis, of the Gov- 
ernment Laboratory of Chicago, suggested to me to increase the amount 
of the reagents used and the length of the time of contact. 

The original method directs that one-thousandth of its weight of pure caustic 
potash should be added to the alcohol and the whole then warmed for an hour at 
a temperature near the boiling point, in a flask under a reflux condenser, and then 
distilled. The alcohol is thus freed from esters and acid. The aldehydes and 
bases are removed by the addition of 10 grams of metaphenylenediamine hydro- 
chloride and 1 gram of phosphoric acid (density 1.4531) to each liter of alcohol. 
The mixture is then warmed under a reflux condenser for one hour at a temperature 
close to the boiling point and afterwards distilled, the first and last 100 cubic 
centimeters being rejected. 

By the use of 10 grams of caustic potash instead of 1 gram, and 
digestion of the solution for one day on a water bath, I secured a more 
complete saponification of the esters and neutralization of the acids; 
and by employing twice the directed amount of metaphenylenediamine 

•Analysis of Potable Spirits. 

^Journ. Am. Chem. Soc. (1906), 28, 397. 

^^Loc cit. 

^^ Loc. cit. 



ALDEHYDES IN DISTILLED LIQUORS. 31 

hydrochloride, allowing the mixture to stand for a week with frequent 
shaking, afterwards digesting it on a water bath under a reflux condenser 
for two days and then distilling, I obtained a sample which when cooled 
to 15° failed to give any color with rosaniline bisulphite. 

Pure aldehyde. — It was prepared by the method of McLeod.^^ This de- 
pends upon the fact that when paraldehyde is heated with diluted sul- 
phuric acid and distilled, aldehyde is obtained almost quantitatively. 
The aldehyde was collected in a receiver with a known quantity of 
water, the wliole, containing 45.06 grams of aldehyde, being diluted to 
200 cubic centimeters. 

Such a mixture keeps for a long time without polymerization. Two 
cubic centimeters of this solution, equivalent to 0.4506 gram of aldehyde, 
dissolved in 500 cubic centimeters of water, was regularly employed in 
this work. 

Reagents. — The reagents are as follows : 

(a) j^ iodine. 

(b) Potassium hydrogen sulphite, about 4 grams dissolved in 100 
cubic centimeters of water. 

(c) Starch paste. 

Determination. — For ordinar}^ work, 25 cubic centimeters of the alde- 
hyde solution (the strength of which must be about 0.5 per cent) are 
mixed with 50 cubic centimeters of the potassium bisulphite in a 150 to 
200 cubic centimeter Erlenmeyer flask. The latter is well stoppered 
and set aside for fifteen minutes, the mixture being occasionally shaken. 
In the meantime, 50 cubic centimeters of the potassium bisulphite are 
titrated by the standard iodine solution, using starch paste as indicator. 
After the expiration of the proper time, the mixture is also titrated with 
the same iodine solution. The difference between the two titrations will 
correspond to the bisulphite combined with the aldehyde in 25 cubic 
centimeters of the mixture. 

Calculation. — Since one molecule of potassium hydrogen sulphite com- 
bines with one of aldehyde, and, on the other hand, to oxidize one 
molecule of potassium hydrogen , sulphite, two atoms of iodine are re- 
quired, the amount of aldehyde can be found from the following pro- 
portion : 

21 ^ M 
I A 

Where 21 represents the iodine combining with one molecule of potas- 
sium bisulphite ; I, the amount of iodine calculated from that of potassium 
bisulphite absorbed by the aldehyde; M the molecular weight of the alde- 
hyde, and A the amount of aldehyde to be determined; hence A = 21* 

"Am. Ghem, Journ. (1907), 38, 27, 



32 



VIVENCIO DEL ROSARIO. 



If n represents the number of cubic centimeters of ^ iodine solution 
used, then:. 

0.0127nM DM 
"~ 2X127 "20,000 

From these constants the author has calculated the following factors: 
Formaldehyde I X 0. 1 1 827 

Acetaldehyde 1X0.17348 

Benzaldehyde I X 0.4 1788 

Vanillin IX 0.59923 

My results are as follows : 

In a preliminary determination 0.0009 gram of aldehyde in 50 cubic 
centimeters of 50 per cent alcohol gave : 



I 

II 
III 
Average, 



Aldehyde. 
0.0008401 
0.0009450 
0.0009450 
0.00091 



Table I. — Determination of known amounts of acetaldehyde. 
BLANK, 20 CUBIC CENTIMETER KHSO3- 42.30 CUBIC CENTIMETER j^ IODINE (0.9533). 



Cubic cen- 
timeters of 
aldehyde 
solution. 


Grams of 

aldehyde 

used. 


Cubic cen- 
timeters of 

iodine jQ- 


Grams of 

aldehyde 

found. 


Percent- 
age. 


5 
10 
15 
20 
25 
30 
35 
40 
45 


.0045 
.0090 
.0135 
. . 0180 
.0225 
.0270 
.0315 
.0360 
.0405 
.0450 


40.15 
38.00 
35.85 
33.80 
31. 50 
29.55 
27.35 
25. 15 
22.95 
20.70 


.00451 
.00903 
. 01355 
. 01784 
. 02268 
.02677 
. 03139 
.03600 
.04063 
. 04536 


100. 22 
100.34 
100. 37 
99.11 
100.80 
99 14 
99.65 
100.00 
100. 24 
100.80 



Table II. 

20 CUBIC CENTIMETER KHSO3- 38.15 CUBIC CENTIMETER ^ (1.0146). 



5 


.0045 


36.15 


.00447 


99.36 


10 


.0090 


84.10 


.00905 


100.61 


15 


.0135 


32.10 


.01353 


100.22 


20 


.0180 


30.15 


. 01788 


99.33 


25 


.0225 


28.00 


.02269 


100.84 


30 


.0270 


26.00 


.02716 


100.59 


35 


.0315 


24.00 


.03163 


100.41 


40 


.0860 


21.90 


.03633 


100.91 


45 


.0405 


19.95 


.04069 


100.46 


50 


.0450 


17.95 


.04516 


100.33 



From these results it is evident that the volumetric method as out- 
lined will undoubtedly be useful to the commercial analytical chemist. 



LARD FROM WILD AND DOMESTIC PHILIPPINE HOGS 
AND THE CHANGES IN THE CONSTANTS PRO- 
DUCED BY FEEDING COPRA CAKE. 



By H. D. GiBBS and F. Agcaoili. 
{From the Laboratory for the Investigation of Foods and Drugs, Bureau of Science.) 



During the examination of lards found upon this market in the search 
for violations of the Food and Drugs Act, some samples came to 
our notice which had remarkably low refractive indices and iodine 
absorption (Hanus) numbers and high saponification (Koettstorfer) 
numbers. This, in connection with the other constants, led to the 
suspicion that coconut oil had been used as an adulterant. No crystals 
characteristic of mixed cholesterol and phytosterol could, however, be 
obtained. It was suggested to us that perhaps coconut cake or coconut 
meal was used as a feed for the hogs and that the constants were thus 
influenced. A claim to this effect was in fact made to us by one firm. 

As a consequence an investigation was undertaken to determine how 
much the lard constants might be affected by feeding copra cake and 
whether a ration containing copra cake could satisfactorily be used here 
for fattening hogs. 

The growing of coconuts is one of the most important industries of 
the Philippine Islands, and during the fiscal year 1909 the quantity 
of copra exported was 105,564,781 kilos valued at $0,657,7-10 dollars 
United States currency. Very little oil is expressed in the Islands. 
The copra cake upon this market contains from 7 to 15 per cent of 
coconut oil. If the climatic and other conditions are such that copra 
cake can be utilized as a food for hogs and cattle in the Philippines, it 
should prove to be a cheap fattening material. While the experiments 
so far conducted indicate that its use in small quantities in the ration 
produced good results in the fattening of hogs, and a profitable use for 
copra cake, further experiments upon hogs and cattle should be conducted 
in the Philippine Islands. 

It is well known that the character of the fats fed to a growing animal 
more or less affects the character of the fat of the animaL A full-grown 
animal on being given a sufficient amount of food after a long period 

93401 3 33 



34 GIBBS AND AGGAOILI. 

of starvation, will deposit a body fat showing some characteristics of the 
fat in the ration. 

Eosenfeld ^ has expressed the opinion that, with natural rations, the 
properties of the fats of carnivorous animals depend for the greater part 
upon the character of the fat eaten, while this is less so for the her- 
bivorous animals, since the latter are better adapted for building their 
body fats from carbohydrates. The ^^omnivorous" hog is, for economic 
reasons, fattened upon carbohydrates. The fat of a grown animal in 
normal condition is not affected by the character of the fat in the ration, 
provided the same normal conditions are maintained.^ For the purpose 
of investigating these points, corn, palm, cotton seed, linseed, and rape 
oils or cakes, some fatty acids, butter, and mutton fats have been 
employed. 

On the other hand, it has been shown that milk fat is, to some extent, 
affected by the character of the fat in a normal ration. The influence 
of coconut oil, in this regard, h^ been investigated to a considerable 
extent. 

Morse ^ found that coconut oil fed to cows changed the iodine number of the 
butter fat from 31.6 to 24.2. 

Siegfeld * observed that coconut-oil cake exerted no apparent influence upon 
the Reichert-Meissel, while the saponification and iodine numbers were influenced 
to a marked extent. 

Knieriem and Buschmann" found that the constants of the butter fat were 
materially affected by feeding cows on rations of coconut cake, linseed cake, rape 
seed cake, etc. 

Tancr§® states that palm-nut cake is an excellent feeding stuff for milch cows 
when fed not to exceed 4 pounds daily and that coconut oil gives similar results. 

The effects of coconut oil upon the fat of the hog have not been investi- 
gated, to our knowledge. Wheat bran as a hog feed, makes a yellow, 
liquid lard and is not suitable. Fish meal makes the lard taste fishy. 
Some lards give the Becchi and Halphen reactions when the hogs have 
been fed on cotton seed cake. 

It has also been shown that the habit and environment will affect 
the fat secreted. For example, the fat of the wild goose has an iodine 
number of 99.6, while that of the bird in captivity is 67."^ Lewkowitsch 

^Magnus, Levy and Meyer, Handbuch d. Bio-chem. d. Mensch. u. d. Tiere 
(1909), 11, 463. 

* Reviews of the literature are given by Magivus, Levy and Meyer, loc cit., 
445, and Emmet and Grindley, Journ. Am. Ghent. Soc. (1905), 27, 263. 
. *Ea!p. Sta, Record 4, (1892-3), 662; N. H. Sta. Bui. 16, (1892), 20. 

*Exp, Sta. Record, 18, (1906-7), 172. 

*Chem. Ztg. Rep. (1907), 31, 287. 

*E(cp. Sta. Record 11, (1899-1900), 1084. 

^Voigtlander, Ztschr. f. ang. Ghem. (1898), 859. 



CHANGES IN CONSTANTS BY FEEDING COPRA CAKE. 35 

gives 76.6 to 84.7, for the iodine number of the fat of the German wild 
boars while we have found for the Philippine wild hog fat 54.8 to 67.8. 
Richardson ^ compares the mast fed or oily hog with the wild boar and 
states that the habits of the two are somewhat similar. 

**They are lean, long-snouted, fast-running animals, rather savage in dis- 
position. Little food is given them, but they are allowed to roam the woods at 
random and, by rooting, to obtain what livelihood they may. In the cooler, their 
carcasses may be easily identified, owing to the fact that even at the freezing- 
point, their fat does not harden, while if the leaf or back fat of such an animal 
be placed in room temperature, the oil runs freely from the tissue." 

He calls attention to the fact that the iodine number of the fat of these 
animal's is about the same as the number given by Lewkositsch for the fat of 
the wild boar. 

The constants for the fat of the oily hog are, however, very different 
from those of the wild hog of the Philippines. (See Table V, p. 42). 

Brandt ^ states that : 

"The best season in the year for collecting the fat and preparing lard is in 
January an^ February, as at that time the fat is of greater consistency than 
during the hot months. * * * Summer fat is therefore richer in olein and 
winter fat in stearin. Feeding and the sanitary condition of the pig exert also 
considerable influence upon the consistency of the fat; hence soft and smeary 
fat is also frequently found in th-e cold season." 

These statements do not seem to apply to the tropical hogs which have 
come under our observation. 

The lards rendered by us from the fat of hogs fed at our request by 
the Bureau of Agriculture were all of good quality and the consistency 
was about the same as that of American and Australian lards shipped 
to this market. It is evident that the copra-cake feed introduces the 
chemical characteristics of coconut oil into the lard. As compared with 
the lard from the corn-fed hogs (fed at the same time) the titre is 2.3 
low, refraction (butyro-scale) 0.6 high, the Koettstorfer number 11,5 
high, and the iodine number 11.7 low. When these latter lards (from 
corn-fed hogs) are compared with the values given by Leach the titre 
is found to be slightly higher, the refraction 5.0 lower, the Koettstorfer 
number 2.0 higher and the iodine number about 10.0 lower on the 
average. 

These variations must be accounted for by the inherent characteristics 
of the hog, the tropical environment and the feed. In all of the animals 
killed by us we have found no oily hogs and in every case the lard has 
been surprisingly firm at room temperature (30° C). No samples were 
liquid at this temperature. 

^Journ. Am. Ghent. 8oc. (1904), 26, 372. 
•Animal and Vegetable Fats and Oils (1896), 109. 



36 



GIBBS AND AGCAOILI. 



EXPERIMENTAL. 

In order to determine the effect of coconut oil upon the fat secreted 
by the hog, a feeding experiment ^^ was conducted with coconut cake as 
a basis of the ration. At the same time a check experiment with corn 
was carried on. The details are as follows: 

Six native hogs, of Chinese origin, from 6 to 8 months old, were divided into 
two lots, No. 1 consisting of 2 hogs and No. 2 of 4 hogs. The feeding was begun 
on March 4, 1908. 



Table I. — Feeding data of Lot No. 1 [corn-fed hogs) .^"^ 



Hog number. 


Initial 

weight 

March 4, 

1908. 


Weight 
June 25. 


j Average 
Gain. gain per 
day. 

Kilos. Kilos. 
26.0 ; 0.23 
28.6 0,25 


5 


Kilos. 
29.0 
38.1 


Kilos. 
55.0 
66.7 


G 




Total 


67.1 


121.7 


54.6 , 0.48 





Number of kilos of corn consumed in 113 days, 308.5. Daily average of corn 
consumed, 2.73. 

Table Ia. — Feeding data of Lot No. 2 {copra-fed hogs) .^^ 



Hog number. 


Initial 

weight 

March 4, 

1908. 


Weight 
June 25. 


1 Average 
Gain, j gain per 
1 dav. 

1 


1^2 


Kilos. 


Kilos. 


Kilos. Kilos. 


2 


31.8 
33.6 
33.2 


63.1 

<X).8 
50 5 


31.3 0.19 
27.2 0.17 


3 


4 __ 


Total 


0..0 ^„.„ 1 „.xu 


98 6 


183 4 


84.8 


0.52 









Number of kilos copra cake consumed in 162 days, 405.5. Daily average of 
cake consumed, 2.50. 

'<» We acknowledge with pleasure our indebtedness to the Bureau of Agriculture 
of the Philippine Islands for conducting the feeding and care of the hogs at 
their experiment station at Alabang. 

" Data furnished by the Bureau of Agriculture. 

" Hog No. 1 died twelve days after the experiment was started. 



CHANGES IN CONSTANTS BY FEEDING COPRA CAKE. 



37 



In addition to the copra cake and corn, both lots were given an equal 
amount of green forage, sorghum, teosinte, velvet bean, or grass. 
Further data were not kept. 

In commenting upon this feeding experiment Dr. H. N. Knight, of 
the Bureau of Agriculture, says : 

"Considerable difficulty was experienced in getting the pigs fed on copra to 
eat a ration of pure copra. This has been observed in other cases. A hog six 
or 8 months old does not take readily to a ration of pure copra tneal, while a 
pig of 2 months will start eating a pure copra ration as readily as a ration of 
pure corn. For this reason the copra-fed hogs did not start putting on flesh 
as soon as did the corn-fed ones, and consequently we were longer in finishing 
the copra-fed lot. 

"My observation on this test, as well as on some others, indicate that it is 
inadvisable to feed a ration of pure copra meal, especially where one is finishing 
aged hogs. It does, however, make a very satisfactory feed when used in com- 
bination with other feed stuffs." 

When the hogs had attained their growth they were carefully marked 
and shipped to Manila where they were slaughtered under our direction 
and supervision and the desired samples of fat at once removed and 
placed in glass bottles. The lards were rendered in a steam bath at 
100° and filtered through paper filters in a steam- jacketed funnel. ' 

The constants were determined in every case in duplicate and in the 
case of the iodine numbers over 60 determinations were, made by three 
different persons with different Hanus solutions. 

The results are condensed in Tables II and Ila. 

Table II. — Constants of lords from corn-fed hogs. 



Kind of lard. 


Tltre. 

41.6 
43.7 

41.6 
43.8 

42.7 
42.7 


Refrac- 
tion, 
butyro 
scale. 

43.5 
42.5 

44.6 
45.5 

46.0 
46.0 


Koett- 

storfer 

numbers. 

196.0 
196.1 

199.0 
198.0 

196.2 
196.1 


Iodine 
number 
of lards, 
Hanus. 

48.8 
47.2 

52.7 
46.7 

49.1 
51.7 


Iodine 
number 
fatty 
acids, 
Hanus. 


Free 

acids 

(asoleic). 


Intestinal fat: 
Hog No. 1 _ 


53.3 
50.3 

53.8 
60.0 

50.4 
52.8 


7.17 
7.75 

9.89 
5.08 

3.72 
8.27 


Hog No 2 „ 


Kidney fat: 

Hog No. 1 


Hog No. 2 

Leaf lard: 

Hog No. 1 


Hog No. 2 . _ 





38 



GIBBS AND AGOAOILI. 



Table IIa. — Constants of lards from copra- fed hogs. 



Intestinal fat: 
Hog No. 1 
Hog No. 2 
Hog No. 3 

Kidney fat: 
Hog No. 1 
Hog No. 2 
Hog No. 3 

Leaf lard : 
Hog No. 1 
Hog No. 2 
Hog No 3 



41.1 


45.5 


41.0 


45.0 


42.3 


46.0 


39.1 


44.5 


39.7 


47.0 


41.0 


45.0 




♦ 


39.4 


44.5 


39.5 


45.0 


40.7 


45.0 



207.4 
204.8 
210.6 

213. 7 
210.4 
210.0 

208. 8 
204. 6 
210. 1 



37.4 


40.2 


0.006 


42. a 


42.9 


0..020 


39.2 


40.3 


0.019 


37.3 


41.4 


0.028 


37.2 


42.0 


0. 032 


35.7 


40.3 


0.028 



35.3 I 
42.5 j 
32.5 I 



42.2 


0.017 


46.2 


0.020 


36.2 


0.011 



Since the Eeichert-Meissel number and the ester equivalent ^^ are in 
a measure chara'cteristic of coconut oil, or at least very different from 
the corresponding lard numbers, these figures were obtained in order to 
determine whether the coconut oil had exerted any influence which could 
be detected in this way. 

The Eeichert-Meissel numbers varied from 2.3 to 0.7 and showed no 
regularity of aiiy interest. The ester numbers varied from 1.0 to 0.4. 
A determination of this factor for a sample of coconut oil gave: alcohol 
solution, 6.7; water solution, 27.5; sum, 34.2. While the free acids, 
as shown in Table -II, are high, it is evident that the differences between 
the constants of these lards and normal lards are not due to this cause. 
Moreover it is to b6 noted that the determination of the free acids was 
made some time after all of the other analytical determinations had 
been completed. We attribute the high acidity in a measure to the fact 
that these samples were, in the beginning, not kept on ice. All other 
samples were constantly on ice except when being sampled for analysis. 

Since the constants of the lards from the corn-fed hogs, especially the 
titre, refractive index, and iodine numbers, showed a remarkable variation 
from the usual lard constants, another series of feeding experiments was 
inaugurated. The data concerning the feeding are so meager and un- 
satisfactory that little or no light is thrown upon the questions under 
consideration. ^^ 



"J. Hanus: Ztschr. f. XJntersuch. d. Nahrungs-u, Genussmittel (1907), 13, 18. 

" Five lots, consisting of four hogs each, were fed on rations composed of 
varying proportions of copra cake, corn, and tiquitiqui (a powder produced in the 
cleaning of rice. It consists of small particles of the grain and a small percentage 
of broken hulls.) The available feeding data furnished to us are tabulated as 
follows : 



CHANGES IN CONSTANTS BY FEEDING COPRA CAKE. 
Feeding data on five lots of hogs fed on copra, corn, and tiquitiqui. 



39 



Lot number, kind, and 
amount of feed used. 


Indi- 
vid- 
ual 
num- 
ber. 

1____ 


Sex. 



M. 
M. 
M. 
F. 


Ini- 
tial. 

Kilos. 
8.1 
8.1 
9.6 
8.1 


Mav 

10,* 
1908. 

Kilos. 
9.6 
3.6 
14.5 
13.6 




Weight 






No- 
vem- 
ber!, 
1908. 

Kilos. 
4.5 


Total 
gain. 

KUos. 


June 

23. 
1908. 


Au- 
gust 1, 
1908. 

Kilos. 
10.5 


Sep- 
tem- 
ber 10, 
1908. • 

Kilos. 
8.6 


Oeto- 
berl, 
1908. 

Kilos. 
6.1 




Kilos. 
10.9 


Lot No. 0; fed on corn; 


J2__-. 
3__.. 
4.__- 




amount fed, 533.5 


16. 2 
18.1 


18.1 
19.9 


16.2 
19.9 


16.8 
19.5 


19.9 
18.6 


6.3 
10.5 


Total i 


33.9 


41.3 










1 




M. 
F. 
M. 
M. 

















1_„- 
2.__. 
3.__. 
4-__- 


8.1 
8.1 
9.1 
6.3 


11.4 

n.4 

12.7 
4.1 


15.2 
15. 2 
18.6 
4.5 

53. 5 


11.8 
19.5 
20.8 


3.6 
25.9 
25.9 








Lot No. 1; fed on fiquiti- 
qui; amount fed, 468. 1_ 


27.2 
25.9 


25.9 
23.6 


17.7 
14.5 













Total 


31.6 


39.6 














f\- 

2.._. 
|3-__. 


M. 
F. 
M. 
M. 












Lot No. 2; fed on corn 
and copra; amount 
fed. 710.8 _ _ _- 


7.7 
10.0 
11.8 
13.2 


9.1 
11.8 
21.3 
25.0 


10.9 
14.0 

32.7 


13.6 
18.1 
34.5 
40.8 


14.5 
11.7 
41.3 

48.0 


15.4 

' 22.2 

43.5 

54.0 


14.1 
22.2 
4L7 
55.4 


6.3 
12.3 
29.9 
42.2 




Total 


42.7 


67,2 


1U7.0 


116.1 


135.1 


133.4 


104.7 




l-_- 

t;: 


F. 
M. 
M. 
M. 


Lot No. 3; fed on copra 
and tiquitiqui; amount 
fed, 765. 3 


14.1 
15.3 
14.5 
30.9 


24.1 
22.7 
20.8 


32.7 
27. 2 
28.0 
83. 6 


40.8 
38.6 
40.8 
50.4 


46.8 
46.8 
49.9 
59.0 


48.1 
47.7 
50.8 
61.3 


54.0 
48.6 
52.6 
62.6 


39.9 
32.3 
38.1 
31.8 




Total 


74.8 




121.5 


170.6 


202. 5 


207.9 


217.8 


142.1 




2____ 
3_— 
4____ 


F. 
M. 
M. 
M. 


Lot No. 4; fed on copra; 
amount fed, 660.9 


13.6 
16.8 
14.0 
18.1 


14.0 
21.3 
25. 
22. 2 


18.1 
35. 
33. 6 
28.0 

114.7 


22. 7 
45.9 
38.1 
36. 3 

143.0 


27. 2 
58.7 
48.6 
45.4 


28.0 
60.4 
49.9 
47.7 


31.3 
65.3 
53. 1 
49.9 


17.6 
48.1 
37.7 
31.8 


Total 


62. 5 


82.5 


179.9 


180.0 


199.6 


185.2 









At the end of the feeding period these hogs were shipped to Manila to be 
slaughtered under the direction of the writers. 

It is greatly to be regretted that en route the hogs became so mixed and the 
labels and markings so confused that one lot could not be distinguished from 
another, and neither could the individual hogs be identified. The value of the 
experiment, in showing the affect of the various kinds of feed upon the lards, 
was entirely lost. However, five of the fattest hogs were selected from which 
samples of fat were taken. The analyses, made before we were informed of the 
meagre character of the <|ata, are tabulated as follows: 



40 



GIBBS AND AGCAOILI. 

Constants of lards from native hogs of known origin. 



Kind of lards. 


Titre. 


Refrac- 
tion, 40°, 
butyro 
scale. 


Koett- 

st'orfer 

number. 


Iodine 
number 
of lards 

(Hanus). 


Iodine 
number 
of fatty 
acids 
(Hanus). 


Free 

acids (as 

oleic). 


Intestinal fat: 

Hog No. 1 


39.2 
42.9 
40.9 
41.7 
40.7 

38.1 
41.5 
39.3 
39.8 
39.1 

38.4 
41.6 
39.6 
39.1 
38.9 


42.0 
^ 45.5 
43.0 
43.0 
45.0 

43.5 
45.5 
42. 5 
43.0 
45.5 

43.0 
45.5 
42.5 
43.0 
45.0 


203.4 
197.7 
191.3 
207. 2 
206.5 

215.1 
200.5 
191.5 
208.8 
211.7 

208.0 
200.4 
192.6 
210.4 
209.3 


35.7 


R7 R 




Hog No. 2 




42.6 i 44 1 


0.28 
0.11 
0.23 
0.49 


Hog No. 3 


32.7 
34.0 
43.4 

37.1 
46.2 
33.5 
31.2 
44.7 

39.2 
35.8 
30.0 
34.9 
44.4 


33.3 
35.7 
43.7 

38.6 
47.3 
35.5 
36.6 

44.8 

40.1 
46.6 
35.5 
35.2 
46.5 


Hog No. 4 


Hog No. 5 . 


Kidney fat: 

Hog No. 1 


Hog No. 2 


0. 20 i 


Hog No. 3 _ 


0.23 
0.42 
0.27 

0.23 
0.23 


Hog No. 4 


Hog No. 5 — 


Leaf lard: 

Hog No. 1 — 


Hog No. 2 


Hog No. 3 


Hog No. 4 


0.14 
0.45 


Hog No. 5 


Maximum 


42.9 
38.1 


45. 5 
42.0 


215.1 
192.6 


46.2 
30.0 


47.3 
33.3 




Minimum 




Average 




40.0 


43.8 


203.7 


37.7 


40.0 









The determination of the free acids was in this case made seven months 
after the other determinations. The lards in the meantime were kept in stoppered 
glass bottles in an ice-box. 

These analyses show very remarkable variations from the usual lard constants. 
For lack of data no conclusions regarding the cause are attempted. 

In order to determine if the native, domestic hogs all showed the same 
peculiarities in their fats^ three samples of lards were obtained at the 
slaughter house from a freshly killed hog of unknown origin^ and a 
composite sample of fat was taken from three different hogs. 

Table III. — Constants of lards from native hogs of unknoxmi origin. 



Kind of fat. 


Titre. 


Refrac- 
tion, 40° 
butyro 
scale. 


Koetts- 

torfer 

number. 


Iodine 
numbers 

of lards 
(Hanus). 


Iodine 

number 

of fatty 

acids 

(Hanus). 


Free 
acid (as 
oleic). 


Intestinal 


40.5 
39.0 
40.1 
40.2 


47.5 
48.5 
49.0 
48.0 


196.5 
197.5 
197.0 
197.0 


56.6 
62.6 
62.7 
55. 2 


57.4 
63.8 
64.4 

57.8 


0.13 
0.34 
0.16 
0.26 


Kidney — _. _ _ 


Leaf - 


Composite sample 



CHANGES IN CONSTANTS BY FEEDING COPRA CAKE. 



41 



With the view of throwing some light on the effect of climate upon 
the fat of the hog, a number of wild hogs ^^ were taken and samples of 
fat prepared. The hogs were found to have very little fat and all that 
was obtained from six of the animals was made into composite samples 
and rendered at once at 100° temperature, and after transporting to the 
laboratory in sealed glass bottles the samples were filtered through paper 
filters in a steam- jacketed funnel. 

Table IV. — Constants of lard from Philippine wild hogs. 



Kind of fat. 



Intestinal 

Kidney 

Leaf 

Lewkowitsch values for 
the wild boar 



Titre. 



(16) 

37.3 

38.7 



Refrac- 
tion, 40°, 
butyro 
scale. 



48.5 
50.0 
49.5 



Koetts- 

torfer 

number. 



197.6 
198.1 



195. 1 



Iodine 
numbers 
of lards 
(Hanus). 



54.9 
67.8 
65.8 



Iodine 

number 

of fatty 

acids 

(Hanus). 



60.1 
75.8 
67.8 



76.6-84.7 



Free 
acid (as 
oleic). 



0.86 
0.28 
0.88 



^^ The wild hogs were secured by Mr. Agcaoili while on a visit to his home. 
He writes as follows concerning the hunt (H. D. G.) : 

HUNTING IN THE PROVINCE OF ILOCOS NORTE. 

It was a fresh, clear morning on the 27th of May, 1909, when my uncle, 
myself, and many others started out on a hunting expedition for the mountains of 
Nagpapalcan, about 20 miles east of the town of Piddig, Ilocos Norte. About 60 
hunters, armed with bolos, nets, spears, and accompanied by dogs, composed the 
hunting party. We arrived at the camp at noon and after lunch we were divided 
by the leader into four parties and started out for the game. Within less than a 
mile from the camp we heard that two deer had already been killed. A short 
distance farther toward the foot of the mountain w^e started three wild hogs. The' 
dogs were upon them very quickly before they could reach their sheltering caves. 
We proceeded on our way to the place where the four columns made their union. 
All the hemp nets w^ere spread in positions designated by the older hunters, 
where the deer and hogs were most likely to run, after which each man took 
the position designated by the leader. All of the dogs were then freed and at 
a given signal from the leader all began to shout and hello vociferously to 
frighten the deer and hogs. Men with spears stood near the nets to transfix any 
animals which might become entangled in the effort to escape from the noise. 
The results of the first run were 17 deer and 2 hogs. 

The next day the hunt was resumed early in the morning in another locality. 
Long before noon 29 deer and 1 hog were killed. After luncheon the homeward 
journey was begun. 

I am much indebted to my uncle Mr. Mariano Aris for his organization and 
leadership of the party which enabled us to secure the required samples of lard. 

*' Sample too small for a determination. 



42 



GIBBS AND AGCAOILI. 



The variation in the constants of the lard of the Philippine wild 
hog from the normal lard values is seen to be rather slight and is chiefly 
noticeable in the low titre. The Koettstorfer number is slightly high. 
The iodine values, while about normal, are decidedly lower than the 
constants given by Lewkowitsch for the wild boar which was a native 
of the temperate zone. 

The remarkable variation in the lards from the Philippine hogs due 
to copra-cake feed and the variation from the normal lards of -the tem- 
perate zone is more clearly seen in the following table. 

Table V. — Comparison of the constants of lard from various sources. 



I 



Corn fed, compiled from Table II. 



Maximum . 
Minimum,. 



Average 

('opra fed, compiled from 
Table I {a. 



Maximum . 
Minimum.. 



Average. 



Native hogs of unknown origin, 
compiled from Table III. 



Maximum . 
Minimum.. 



Average 

Values given I 

Maximum 

Minimum 

Values given I 

Maximum ___ 



I Leach. ^^ 



' Lewkowitsch.^^ 



Minimum 

Philippine wild hogs, compiled 
from Table IV. 



Maximum — 
Minimum 



Average 

Lewskowitsch's values. 



Constants for the fat of oUy hogs 
compiled from Richardson, 21 



Maximum . 
Minimum.. 



Titre. 



Reirac- 

tionat 

,40°, buty 

I ro scale. 



43.8 
41.6 



40. 5 
39.0 



39.9 



42.0 
41.0 



4().0 
42.5 



Koetts- 
torfer 
number. 



Iodine 
number 

(Hanus). 



Iodine I 
number 



(Hanus). 



199.0 I 
196.0 ; 



52.7 : 
46.7 [ 



53.8 
50.0 



49.4 



51.7 



47.0 
44.5 



45.3 



213. 7 
204.6 



49.0 
47.5 



48.3 



197.5 
196.5 



51.2 
48.6 



i»53.0 
44.8 



38.7 
37.3 



38.0 



30.8 
24.7 



50.0 

48.5 



49.3 



57.1 
54.0 



196.6 
195.2 



198.1 
197.6 



192.5 
189.0 



42.5 I 
32.5 j 



46. 2 
36.2 



62.7 
55.2 



59. 4 



70.0 
50.0 



2" 70. 4 
49.9 



67.8 
54.8 



62.8 



76.6—84.7 



95.2 
78.8 



64.4 
57.4 



60.9 



75.8 
60.1 



67.9 



"Food Inspection and Analysis. John Wiley & Sons (1904), 410. 

"Oele, Fette und Wachse. Braunschweig (1906), 368. 

"Fat from foot. 

* North American lard from head of hog. 

^^Journ. Am. Chem, 8oc. (1904), 26, 373, and (1908), 30, 1191. 



106.0 
90.0 



CHANGES IN CONSTANTS BY FEEDING COPRA CAKE. 43 

P. Voigtlander 2^ in comparing the American lard with the German 
states that the higher iodine number of the former is explained by its 
greater content of lard oil and the composition of the lard oil. The 
American lard oil has an iodine number of about 88, while the German 
has one between 70 and 75. 71.7 per cent of the lards examined in the 
i'abrik von Dietrich in Helfenberg had iodine numbers between 48 and 
53. Of the American lards, 88.5 per cent examined had iodine numbers 
between 61 and 66, and 41 per cent were over G4. The iodine numbers 
for the fatty acids were found to be between 90 and 95 for German lards 
and between 98 and 104 for American lards. 

From the results of this work it is evident that lean hogs will deposit 
a body fat with constants considerably influenced by coconut oil, when 
copra cake forms only a part of the ration. When the ration is composed 
almost entirely of copra cake the influence upon the constants is more 
remarkabl'3. 

The influence of the tropical climate and that of the inherent charac- 
teristics of the tropical hog upon the lard can not be sufficiently discussed 
from the data at hand. It can be said, however, that the general 
tendency of the hog grown in these environments is to deposit a firm 
body fat having a high titre, high saponification number, and low refrac- 
tion and iodine numbers. 

One of us (H. D. G.) is about to visit the United States. Upon 
returning to the Philippines another series of feeding experiments will 
be attempted, time and press of other work permitting, under our per- 
sonal daily supervision. It is proposed to investigate the questions 
involved with a view of obtaining . sufficient data so that some accurate 
conclusions may be drawn concerning the effect of coconut fat upon the 
body fat of both foreign and native hogs. The investigation will, if 
possible, be extended with the view of determining the effect of the 
tropical environment not only on the fat of the hog but also upon the 
fat of other animals. 

^Loc. cit. 



EDITORIAL. 



THE LIBRARY OF THE BUREAU OF SCIENCE. 

From the very first suggestion of a central laboratory for the several 
Bureaus and Departments of the Philippine Insular Government, early 
in 1901, to the latest report of the United States Philippine Commission, 
the value of the scientific library as a most important part of that institu- 
tion has been constantly emphasized. 

On the appointment of a Superintendent of Government Laboratories 
on June 20, 1901, this official was directed, among other things, to obtain 
information not already on hand, to be used for the equipment and 
library of the Philippine laboratory. In accordance with these instruc- 
tions, lists of journals and manuals absolutely necessary for the beginning 
of a scientific library were compiled before the superintendent sailed from 
the United States on August 21, 1901. 

On the legislative side. Act ISTo. 156, passed by the United States 
Philippine Commission July 1, 1901, gave the Superintendent of Gov- 
ernment Laboratories charge over the purchase of books authorized for 
use in connection with all Government laboratories for the Philippine 
Islands. He was also directed to prepare and recommend, as a basis 
for appropriation by the central legislative body of the Philippines, plans 
for a suitable building for the laboratories and for a reference library, 
together with estimates of the cost of properly equipping the said labor- 
atories and of procuring an adequate reference library. As a result of 
these recommendations, the governing body agreed to appropriate, from 
time to time, in amounts best suited to existing conditions, the sum of 
$46,290.66 United States currency for the establishment of this library. 

On the arrival of the Superintendent of Government Laboratories in 
Manila on September 25, 1901, considerable time was spent in securing 
a temporary building, inaugurating the most urgent lines of work, 
cabling for bids on books and apparatus, and preparing plans for the 
permanent building, but notwithstanding this delay, a generous order for 
subscriptions to serial publications and for the most essential manuals 
was sent out early in 1902, and before the end of that year a very creditable 
showing was made on the shelves of the various rooms in the temporary 
laboratory building. Early in 1903, it was recognized that this material 
was accumulating so rapidly that a separate room was necessary for it 
and that more care than the stenographers and laborers of the Bureau 
could give it, was essential. Accordingly, a room was built and all books 

45 



46 EDITORIAL. 

and journals, then amounting to some two hundred subscriptions and 
probably one thousand manuals and sets of periodicals, were transferred 
to the closed cases installed therein and the writer was appointed librarian. 
Even before that date, the library had been the recipient of , a large 
number of documents of the United States and foreign governments and 
of a considerable amount of material in the form of reprints and separate 
monographs on scientific subjects. From that time to the present, addi- 
tions of similar publications as gifts and as exchanges have contributed 
in no small measure to the value of the library. 

It was expected that the building provided for by Act No. 156 would 
be ready for use by July 1, 1904, but long before this the library so far 
outgrew its restricted quarters that a larger room was assigned to it 
in a dwelling which was rented early in 1904 for the accommodation of 
a part of the laboratory work. Here the books were shelved in open 
wooden cases, the legs of which were placed in petroleum cans as a pro- 
tection against the attacks of white ants, with which this building was 
infested. 

Before the end of August, 1904, this space had become so crowded, 
that in order to locate a new periodical set in place on the shelves, it 
was often necessary to move all the books in from eight to ten cases. 
It was, therefore, a great satisfaction to change from these crowded 
quarters to the present library rooms in the Laboratory Building on 
Calle Herran on September 1 of that year. Here provision had been 
made for shelving from 30,000 to 40,000 volumes, two stack rooms each 
24 by 16 feet having been added after the original plans for the building 
were drawn. However, the growth of .the library since 1904 has been 
so rapid that the limits fixed in 1901 have now been reached and the 
need for more room is urgent. It is hoped that this may be met by 
the construction of a fireproof building, equipped with standard library 
furnishings of all kinds and provided with the best modern lighting 
and ventilating systems. 

The books in the Calle Herran building were placed upon open shelv- 
ing of a standard adjustable type, with metal stacks and wooden shelves. 
These have proved satisfactory, although constant supervision is neces- 
sary to prevent rusting of the metal brackets and screws. 

Aside from adequate shelving, the care of books in the Tropics presents 
many problems unknown in a temperate climate. Constant attention is 
necessary to prevent them from being damaged by insects and mold. 
Fortunately, the library has as yet suffered no serious loss from these 
enemies, but the methods of combating them are still experimental, hence 
it is not certain that the books are free from danger. 

Act !N'o. 156 providing only for a biological laboratory, including 
pathology and medicine, and a chemical laboratory, therefore it was 
natural that these two divisions of the library should have been em- 



EDITORIAL. 47 

phasized from the beginning and that they should lead in the number 
of volumes and parts on the shelves at the present time. As the work 
of the laboratory developed along other lines, demands for literature 
on many new subjects arose. With the transfer of the Government 
botanist, who had previously been connected with both the Bureau of 
Agriculture and the Bureau of Forestry, a number of valuable works 
were added and orders were placed for many more. The demand for 
systematic and economic investigations of the insect life of the Islands 
led to the appointment of an entomologist and for this work an effort 
was made to collect the literature on this subject, especially that relating 
to the Philippines, which had appeared in so many different publications. 
Xext came the organization of the Serum Institute, and with it a great 
demand for works on veterinary subjects and on immunity and serum 
therapy. Later a natural history collector was transferred from the 
former Philippine Museum and for the. proper identification and descrip- 
tion of his collections many important works in general zoology, mam- 
malogy, and ornithology were needed. 

On October 20, 1904, the Secretary of the Interior, at a meeting of 
the Bureau Chiefs of his Department, proposed a plan for collecting all 
books and periodicals belonging to the different Bureaus of the Depart- 
ment of the Interior into one general scientific library, property re- 
sponsibility to rest with the central library, all orders to be sent from it, 
and every effort made to keep periodical files complete for the various' 
Bureaus, they to withdraw books needed for regular wprk from the 
central library on memorandum receipt. By this plan, it was hoped 
that unnecessary duplication in the purchase of expensive works would 
be avoided, that information concerning the location of any particular 
scientific publication would be readily available, that, with a minimum 
of effort, all volumes of periodicals and serials could be completed and 
bound promptly, and that eventually a central catalogue might give 
necessary information concerning all works, while duplicate sets of 
catalogue cards for books retained in the various Bureaus could be fur- 
nished, at little expense, with each collection withdrawn, and finally, 
that more money might become available for the purchase of scientific 
books. Much of this has been accomplished, but a great deal remains to 
be done. The check on duplication apparently has been considered 
successful, since this point was one emphasized in the establishment of 
the Philippines Library Board by Act No. 1935, effective May 20, 1909. 
A large number of volumes have been bound for 'other Bureaus, and 
information furnished in many instances concerning the location of 
particular scientific works. 

Under this arrangement about 6,800 volumes were added to the central, 
library. In order to have a record of this new material and to convey 
to interested persons some idea of the resources of the scientific collec- 



48 



EDITORIAL. 



tions in Manila, a list of all titles (by author entry) was issued as Part 
II of Publications of the Bureau of Government Laboratories, No. 22. 
This list appeared early in 1905 and contained all additions to the 
library up to January 31 of that year. 

At the close of the Louisiana Purchase Exposition in St. Louis, much 
material from the Philippine exhibit was returned to Manila. From 
this collection during the summer of 1905, the Bureau of Science received 
some 250 volumes, including a number of extremely valuable w^orks, such 
as Sempe/s Reisen im Archipel der Philippinen, 1868-1902, etc., and a 
large number of the more important smaller contributions to Philippine 
anthropology and ethnology. 

On November 1, 1905, by provision of Act No. 1407, commonly re- 
ferred to as the Reorganization Act, the former Mining Bureau became 
a division of the new Bureau of Science created by the same act, and 
all books, periodicals, maps and catalogues were transferred to the shelves 
of the central library. The former Ethnological Survey, by the same 
Act, was created a division of the Bureau of Education and about a 
year later. was transferred to the Bureau of Science, bringing to the 
shelves of the central library not only the volumes held on memorandum 
receipt, but a considerable amount of material which had been collected 
in the preliminary organization preparatory to the establishment of a 
Commercial Museum. 

The inauguration of the marine biological work, of the division for 
the standardization of weights and measures, and of the food and drug 
laboratory have brought large additions to the library during tlie past 
three years. 

From the above it will be seen that the library has not only inci eased 
rapidly in the number of volumes, but also in the extension of classes 
represented. The following table gives the number of bound and un- 
bound volumes and parts classified and shelf -listed prior to January 1, 
1910, together with one hundred titles representative of the best material 
in the larger sections : 



Classes and titles. 



Qeneral works 

History (except America) 

History (America) 

Geography— Anthropology 

The Geographical journal. 26 v. 

Petermann's mitteilungen. 53 v. 

Archiv ftir anthropologic. 34 v. in 31. 

Royal anthropological institute of Great Britain and Ire- 
land. Journal. 33 v. 

Soci4t^ d' anthropologic. Bulletins et memoires. 47 v. and 6 
no. in 5. 



Bound 


Unbound 


Parts. 


volumes. 


volumes. 


187 


67 


1 


36 


19 




27 


6 


50 


353 


96 


34 



EDITORIAL. 



49 



Glasses and titles. 



Bound 
volumes. 



Unbound 
volumes. 



Parts. 



175 V. 



Economics 

Political science 

Education , 

Literature 

General science 

The American journal of science. 

Nature. 76 v. 

Smithsonian institution. Annual report of the Board of 

regents. 57 v. Contributions to knowledge. 29 v. and 12 

nos. Miscellaneous collections. 34 v. and 37 nos. 
U. S. National museum. Report. 28 v. Proceedings. 35 v. 

Bulletin. 68 v. in 48. 
Royal society of London. Philosophical transactions. 164 v. 

in 194. Proceedings. 82 v. in 81. 
The London, Edinburgh, and Dublin philosophical magazine 

and journal of science. 231 v. in 192. 
K. Akademie der wissenschaften. Vienna. Mathematisch-na- 

turwissenschaftliche klaasc. Sitzungsberichte. 115 v. in 193. 
Academic des sciences. Paris. Comptes rendus. 145 v. 
K. Preussische akademie der wissenschaften. Berlin. Mo- 

natsberichte. 31 v. Sitzungsberichte. 52 v. 
Asiatic society of Bengal. Journal and proceedings, 43 v. 

in 24. 
Royal Asiatic Society. Straits branch. Singapore. Journal. 

53 V. 
Tokyo. Imperial university. College of science. Journal. 

24 V. 
Australian association for the advancement of science. Re- 
port. 10 V. 
Natuurkundig tijdschrift voor Nederlandsch Indie. 67 v. 

in 65. 
Challenger expedition. Report on the scientific results of the 

voyage of H. M. S. Challenger. 40 v. in 50. 

Mathematics 

Astronomy 

Physics 

Chemistry 

The Analyst. 32 v. in 31. 

American chemical journal. 40 v. 

American chemical society. Journal. 29 v. in 23. 

Annales de chimie et de physique. 372 v. 

Chemical society. London. Journal. 95 v. in 117. 

Die Chemische Industrie. 30 v. 

Chemisches central-blatt. 75 v. in 100. 

Deutsche chemische gesellschaft. Berichte. 41 v. in 117. 

Jahresbericht iiber die fortschritte der chemie und verwandter 

theile anderer wissenschaften. 59 v. in 83. 
Journal fiir praktische chemie. 253 v. 
Justus Liebig's Annalen der chemie. 369 v. 
Soci6t6 chimique de France. Bulletin. 88 v. 
Society of chemical industry. London. Journal. 27 v. 
Zeitschrift fiir angewandte chemie. 20 v. in 24. 
Zeitschrift fiir physikalische chemie. 65 v. 

93401 4 



71 
42 
8 
6 
2,107 



266 
1* 
28 
2 

679 



8A 
2 



401 



17 


8 


16 


7 


405 


64 


2,386 


152 



48 



50 



EDITORIAL. 



Classes and titles. 



Geology 

Congr^s g6ologique international. Comptes rendus. 13 v. 

Geological society of America. Bulletin. 18 v. 

The Journal of geology. 15 v. 

U. S. Geological survey. Annual report of the director. 63 v. 
in 68. Bulletin. 376 v. Monographs. 46 v. in 48. Pro- 
fessional papers. 64 v in 66. Mineral resources of the 
United States. 18 v. Water-supply papers. 231 v. 

Spain. Comisi6u del mapa geol6gico. Boletin. 21 v. Me- 
morias. 18 v. in 17. 

Pal aeon tographica. 54 v. in 56. 
Natural history 

The Annals and magazine of natural history. 141 v. in 71. 

Quarterly journal of microscopical science. 59 v. 

Royal microscopical society. London. Journal. 30 v. 

Archives de biologic. 23 v. 

Biometrika. 5. v. 

Zeitschrift fur biologic. 49 v. 
Botany ^ 

Annals of botany. 21 v. 

Botanische jahrbiicher fiir systematik, pflanzengeschichte 
und pflanzengeographie. 9 v. 

Deutsche botanische gesellschaft. Berlin. Berichte. 25 v. 

Jahrbiicher fur wissenschaftliche botanik, 45 v. 

The Journal of botany, British and foreign. 45 v. 

Just's Botanischer jahresbericht. 34 v. 

Kew. Royal botanic gardens. Bulletin of miscellaneous 
information. 28 v. in 18. 

Linnean society. London. Journal. Botany. 37 v. in 33. 
Transactions. 2d series. Botany. 6 v. 
Zoology 

The Zoological record. 44 v. 

Zoological society of London. Proceedings. 78 v. in 58. 

Zoologischer anzeiger. 32 v. in 30. Bibliographia zoologica. 
14 V. 
Entomology 

Berliner entomologische Zeitschrift. 52 v. in 36. 

Entomological society of London. Transactions. 55 v. 

Soci6t6 entomologique de Belgique. Brussels. Annales. 51 
V. in 44. 

Soci6t6 entomologique de France. Paris. Annales. 77 v. in 
78. Bulletin. 20 v. 
Human anatomy 

Archiv fur mikroskopische anatomic und entwicklungsge- 
schichte. 73 v. 

Journal of anatomy and physiology. 43 v. 
Physiology 

Hoppe-Seyler's Zeitschrift fiir physiologische chemie. 60 v. 
Bacteriology 

Centralblatt fiir bakteriologie, parasitenkunde und infek- 
tionskrankheiten. 79 v. 

Jahresbericht iiber die fortschritte in der lehre von den patho- 
genen mikroorganismen. 23 v. 

Paris. Institut Pasteur. Annales. 22 v. Bulletin. 6 v. 



Bound 
volumes. 



429 



1,345 



Unbound 
volumes. 



816 



112 



320 



778 



248 



Parts. 



210 



284 



53 



EDITOBIAIj. 



51 



Classes and titles. 



Bound 
volumes. 



Unbound 
volumes. 



Parts, 



General medicine __ 

The American journal of the medical sciences. 161 v. 

The Journal of experimental medicine. 10 v. 

The Journal of medical research. 19 v. 

The British medical journal. 103 v. 

The Lancet. 175 v. in 171. 

Archives de medicine navale. 9 v. 

Berliner klinische wochenschrift. 44 v. in 45 and 49 no. 

Deutsche medizinische wochenschrift. 34 v. in 38. 

Miinchener medizinische wochenschrift. 56 v. in 61. 

Schmidt's Jahrbiicher der in- und auslandischen gesammten 
medicin. 302 v. in 238. 

Virchow's Archiv fiir pathologische anatomic und physiologic 
und fiir klinische medizin. 195 v. in 203. 
Hygiene, etc 

Zeitschrift fiir hygiene und infektionskrankheiten. 62 v. 
Pathology 

Archiv fiir experimentelle pathologic und pharmakologie. 59 
V. in 36. 

Beitrage zur pathologischen anatomie und zur allgemeinen 
pathologic. 43 v. 

The Journal of pathology and bacteriology. 13 v. 
Practice of medicine 

The Journal of infectious diseases. 4 v. 

Archiv fiir schiffs-und tropen-hygiene. 11 v. 

The Journal of tropical medicine and hygiene. 9 v. 

Liverpool school of tropical medicine. Memoir. 21 v. 
Thompson Yates and Johnston laboratories. Report. 7 v. 
in II. 

Surgery 

Ophthalmology — 

Otology 

Gynecology and obstetrics 

Pediatrics 

Dentistry 

Dermatology 

Therapeutics 

Pharmacy and materia medica 

Archiv der pharmazie. 194 v. in 111. 
Agriculture 

Agricultural bulletin of the Straits and Federated Malay 
States. 7 v. 

The Tropical agriculturist. 30 v. 

U. S. Department of agriculture. Annual report. 11 v. Year- 
book. 15 v. Farmers' bulletin. 399 nos. , etc. etc. 

Biedermann's Central-blatt fiir agrikultur-chemie. 36 v. 

Jahresbericht iiber die fortschritte auf dem gesamtgebiete der 
agrikultur-chemie. 51 v. in 43. 

Plant culture 

Forestry 

The Indian forester. 17 v. and 11 nos. 
Animal culture and veterinary medicine 

The Journal of comparative pathology and therapeutics. 20 v. 

Journal de m^decine v6t6rinaire et de zootechnie. 58 v. 



8,455 



63 



660 



350 
252 



198 



40 



114 



105 

24 

5 

51 



175 
27 
289 



11 



1 

11 
4 

255 



104 
59 



312 



242 
207 



25 



52 



EDITORIAL. 



Classes and titles. 



Bound Unbound 
volumes, volumes. 



Parts. 



Fish culture 

Technology 

Dingler's polytechnlsches journal. 321 v. 

Engineering 

Roads and pavements 

Mineral Industries 

American institute of mining engineers, New York. Transac- 
tions. 35 V. 

The Mineral industry. 17 v. 
Chemical technology 

Jahresbericht iiber die leistungen der chemischen technologic, 
55 V. in 60. 

Manufactures 

Foods 

Bibliography 



37 
366 

48 
10 
138 



40 


7 


47 


3 


137 


187 



Besides the sections noted above^ there is still a large amount of 
unclassified material in geology^ mining and mineral industries, technol- 
ogy, geography and anthropology, travel, history, photography, fish and 
fisheries, political and social science and bibliography. If we add to the 
above, reprints, dissertations and monographs. Government publications 
not bearing directly upon the work of the Bureau, incomplete series of 
publications not considered of sufficient value to complete and bind, and 
duplicate material, it will easily be seen that we have probably already 
gone beyond the outside figures suggested in planning for the future 
growth of the library in its present quarters. 

During the first five years almost the whole time of the small library 
staff was given to securing publications, and completing and building sets 
of periodicals, with only such temporary records as would enable works 
to be found on the shelves. From the beginning, the books were ar- 
ranged roughly according to one of the standard classification schemes 
and arbitrary symbols assigned for locating them. The transition to 
the present arrangement on the shelves was, therefore, not sudden, so 
far as the main classes were concerned, but in a library growing as 
rapidly as this one from 1903 to 1908, it frequently happened that the 
actual location of a given book might change a great many times during 
a year, although the relative position remained the same. If we add to 
this fact the one that books bore no key to their relative locations, the 
problem of keeping them on their proper shelves can easily be seen. 
The accession book was first brought up to date, and the cost of the 
library was segregated from that of apparatus and supplies of the Bureau. 
Because of clerical errors in volume numbers, dates, etc., and of the fact 
that well-meaning people along the way had frequently attempted to 
assist us by translating several titles of foreign journals by the same 



EDITORIAL. 53 

English equivalent^ thus completely losing the identity of the originals, 
this latter was a slow and nerve-wearing process. However, by January, 
1908, most of these matters were fairly worked out, and since that date 
the strongest emphasis has been placed upon the work of permanent 
organization. In this, we have been fortunate in having an assistant 
librarian with long library experience and with recent training in the 
shelf-listing division of the Library of Congress. The shelf-list has 
been pushed ahead of the catalogue, and on January 1, 1910, a total of 
25,837 volumes and 2,276 unbound parts had been classified, shelf -listed, 
labeled and placed in permanent position. In addition to the- official 
shelf-list, a duplicate, differing from the original only in the omission 
of accession numbers, has been placed in the reading room, which, with 
an author index, makes it possible to find any classified work very 
promptly. 

Although the classification and shelf-listing have frequently been 
interrupted by the demands of routine work and, by change of clerks 
assisting in typewriting, filing, etc., the work has progressed as rapidly 
as could have been hoped for, and the time saved in library administra- 
tion by the aids now available for use more than covers the additional 
cost of the work to date. 

The system of classification, decided upon after careful consideration, 
is the one in use in the Library of Congress. This system has proved 
especially well adapted to science, medicine and public documents, of 
which at least 90 per cent of this library at present consists. 

Mary Polk. 



REVIEWS. 



light and Sound. By William S. Franklin and Barry MacNutt. Cloth. Pp. 
vii-344. Price, $1.60 net. New York: The MacMillan Company. London: 
MacMillan & Co., 1909. 

This book is the third one of a series written by these authors on 
special subjects in physics. It professes to be a text-book for colleges 
and technical schools. It is evident that the primary object of the 
authors has been to write a popular book. Many things which seem 
vital to the reviewer have been sacrificed by the authors in an attempt 
to make the subject interesting. 

The subject is treated in a very elementary manner and with the use 
of very little, mathematics. As physics is essentially a mathematical 
science, the reviewer believes there will never be in the future any great 
physicst who will not be able to think in the terms of mathematics ; and 
even a beginning course in physics should lay the foundation for 
mathematical thinking. 

The authors state that the main object of the book is to develop "a 
simple living interest in science on the part of all plain people.^^ As a 
popular and accurate reading book of selected practical chapters on 
Light and Sound, the volume is excellent. A text book of physics should 
develop the subject from a mathematical point of view, and if such a 
treatment is dry and lacks interest, the latter must be developed through 
the lectures and discussions of the class room and especially through the 
enthusiasm and personality of the teacher. 

The subjects selected cover only such parts of the fields of light and 
sound as are commonly met with in every day experience and have 
evidently been so chosen by the authors because of the greater interest 
and ease of treatment by practical illustrations. There is only an exceed- 
ingly limited discussion of recent theories of light, and after stating that 
the electric magnetic theory of light is the one generally accepted by scien- 
tists, the authors, for reasons of simplicity, in most parts of the book use 
the elastic solid theory in their explanations of light phenomena. 

There are ten chapters on light. The wave theory as applied to light 
is illustrated and made quite simple by the use of water waves. The 
chapters on Lenses and Lens Systems, Simple Optical Instruments, and 
Lens Imperfections and their Compensation, are three of the most prac- 
tical in the book. They explain in a simple manner the various phenom- 

55 



56 REVIEWS. 

ena of light which enable the reader intelligently to understand lenses 
and their uses in optical instruments. 

Methods of measuring light and the standards used in these measure- 
ments are concisely stated in the chapter on Photometry and Illumination. 
Many suggestions are given in this chapter for the proper illumination 
of rooms, halls, etc. The chapters on Eeflection and Eefraction, Dis- 
persion and Spectrum Analysis, Interference and Diffraction, Polariza- 
tion and Double Eefraction, are dealt with in their logical order in a brief 
and satisfactory manner. 

The last six chapters of the book are on Sound. As the author^s state- 
ment clearly indicates, the treatment of the subject is "abbreviated to 
the utmost and the discussion is limited to those things which are of 
practical importance.^^ The chapters, The Physical Theory of Music, 
and Miscellaneous Phenomena Depending upon the Eeflection, Eefraction 
and Diffraction of Sound and Architectural Acoustics, are interesting 
and practical. 

The book would seem to fill the authors' ideal of making the subject 
interesting, but to the mind of the reviewer it would serve better as an 
auxiliary reading book in physics than as a text-book. There are many 
references, principally to the standard books on physics, given in the 
text. The typography, illustrations and printing are good. A poor 
quality of paper, however, has been used. 

E. F, B. 



An Introduction to Chemical Analysis for Students of Medicine, Pharmacy and 
Dentistry. By Elbert W. Rockwood, M. D., Ph. D. Third edition, revised, 
with 20 illustrations. Cloth. Pp. xiii-247. Price, $1.50. Philadelphia: P. 
Blakiston^s Son & Co. 

The greater prominence given to the ionic explanations of analytical 
reactions is practically the only change over the preceding editions and 
constitutes the most commendable feature of the text. 

It is the belief of the reviewer that elementary instiiiction in quali- 
tative volumetric analysis has no ligitimate place in the professional 
schools of medicine, pharmacy, or dentistry. 

The subjects under Applied Analysis, viz, Waters and Blowpipe 
Analysis and Detection of Poisons, are well chosen, but are altogethei 
too briefly treated to be of great value. 

G. F. K. 



o 



0. 



< 




HULBURT AND BACON : A LIQUID CONCAVE MiRROR.] [PHIL. JoURN. SCI., VOL. V, No. 1. 




PLATE II. PHOTOGRAPH OF FILM OF AN INCANDESCENT LIGHT BULB. 



• . . :•• 



THE PHILIPPINE 

Journal of Science 

A. Chemical and Geological Sciences 
AND THE Industries 



Vol. Y 



MARCH, 1910 



No. 2 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN 

LUZON. 



By Geobge I, Adams. 
{From the Division of Mines, Bureau of Science, Manila, P. I. 



CONTENTS. 



Introduction. 
Physiographic regions. 
Western cordillera. 
Central plain. 
Eastern cordillera. 
Loboo Mountains. 
Southwestern volcanic region. 
Geology. 

Western cordillera. 
Geologic mapping. 
Older sedimentaries. 
Igneous rocks. 
Limestones. 
Marine conglomerates. 
Tufaceous foraminiferal marls. 
Geologic history. 
Central pla,in. 

. Water-laid tuff. 
• Alluvium. 
Eastern cordillera. 
Schists. 
Diorites. 
Andesites. 
Basalts. 
93217 



Geology — Continued. 

. Eastern cordillera — Continued. 

Tertiary limestones, sandstones 
shales, and coal. 
Loboo Mountains. 

Diorities. 

Andesites. 

Tertiary sedimentaries, 

Coralline limestones and marls. 

Geologic history. 
Southwestern volcanic region. 

Extinct volcanoes. 

Eniptives. 

Cinder cones. 

Crater lakes. 

Active Volcano Taal. 

Historic eruptions of Taal. 

Agglomerates. 
, Tuff formation. 

Coralline limestones i^nd marls. 

Classes of igneous rocks. 

Structure of the region. 

Origin of Taal Lake. 

Origin of Bay Lake. 

57 



58 ADAMS. 

CONTENTS — continued. 



Mineral besources. 
Iron. 

Lanatin iron deposits. 
Iron deposits near Santa Inez. 
Copper. 

Prospects in the Loboo Moun- 
tains. 
Prospects on the Lanatin River. 



^«ld- i Sand. 



Mineral resources — Continued. 
Stone. 

Guadalupe and Meycauayan 

stone. 
Sisiman andesite. 
Basalt. 
Gabbro. 
Limestone. 
Gravel. 



Coal. 

Bulacan coal field. 
Gas. 



Clay and clay products. 

Salt. 

Lime. 



INTRODUCTION. 



The area discussed in this report has been visited in part by nearly all 
of the geologists who have studied the Philippine Islands since it includes 
Manila, the capital and principal port, and a convenient starting point 
for excursions. However, no systematic reconnaissance has been made 
previous to this report, and the literature, which is in several languages, 
is largely fragmentary and not readily accessible. 

The most important geologic map published is by Yon Drasche, who 
saw a considerable part of the country and drew upon the writings of 
others and probably used data furnished him by residents of the island. 
In addition there is a small map of Mount Maquiling by Abella and a 
partial mapping of the tuff deposits by Centeno. 

The present reconnaissance was planned so as to make possible the 
correlations of the many observations of previous workers and the pub- 
lishing of a map to serve as a guide to the understanding of the broader 
geologic problems. The localities described by former writers have been 
revisited with the exception of some lately studied by the members of the 
present division of mines. 

It is to be regretted that the uninhabited and heavily wooded portion 
of the eastern cordillera must be left unexplored, and that many in- 
teresting areas could not be studied more in detail because the general 
problems required all the allotted time. 

Aside, from the purely scientific interest, the reasons for making this 
reconnaissance were the need for more accurate knowledge in regard to 
the coal deposits near Manila, the availability of structural materials 
required in the economic* development of the island, and a better under- 
standing of the geologic conditions governing the occurrence of artesian 
and deep-seated waters which are so necessary for the well-being of the 
inhabitants. Large sums of money are being spent for public improve- 
ments in this area by the Government under the direction of the Bureau 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 59 

of Public Works, by the Army Engineers, and by the municipalities, and 
it is desirable that the division of mines, when called upon, should be 
able to supply data which will aid in the various undertakings. With 
this purpose in mind, practical results have already been reached along 
certain lines ^ and investigations are being carried forward which can 
not be completed in time to contribute to this paper. 

The region studied does not contain any developed metallic resources, 
and in the part explored it is questionable if any exist, but the value 
of the structural materials used during the past year exceeds the present 
output of the best developed mining districts of the Islands an*d the 
value of the underground waters brought to light can hardly be estimated. 
It is probable that the coal fields in Bulacan will soon be thoroughly 
investigated. 

The difficulties which attend geologic explorations in the Philippines 
even in this comparatively accessible region will not be enumerated. 
They have been mentioned by many writers, and those who know them 
will excuse certain of the weaknesses of this report. 

Mr. Eobert N. Clark was my field aid and traveling companion, and 
I take this opportunity to acknowledge his valuable assistance and 
cheerful cooperation. 

PHYSIOGRAPHIC REGIONS. 

On the index map, fig. 1, the general physiographic regions of Luzon 
Island are outlined for the first time. It is to be regretted that these 
regions are not well enough known to warrant description. The rectangle 
defined by heavy lines indicates the area discussed in this report and the 
geography of it will be seen more in detail on the accompanying geologic 
map (folded Map 1). 

Western cordillera, — The southern portion of this region, also called 
the cordillera of Zambales, as seen from the east, presents two gaps 
defining two mountain masses near the end of the range. The southern 
mass is known as the Mariveles Mountains, with Mount Mariveles having 
an elevation of 1,419 meters. The second contains Mount Natib which 
has an elevation of 1,285 meters. To the north of the second and 
higher gap are Mount Caulaman, elevation 928 meters, and Mount 
Dangas, elevation 938 meters. Mount Pinatubo is reached by continuing 
about 20 kilometers beyond the limits of the geologic map. Its comb- 
like peak, which has an elevation of 1,806 meters, is considered the 
highest point in the cordillera. 

From the indenture of Subig Bay, on which the naval station of 
Olongapo is situated, one may travel by a low. gap to the wide alluvial 
valley occupied by the Pinatuan and Santo Tomas Kivers which flow 
into the China Sea. These features separate the Cinco Picos Mountains 

"■This Journal, Sec. A (1909), 4, 455 and 463. 



60 



ADAMS. 




Pig. 1. — iNDBx Map Showing Physiographic Regions of Luzon. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. (51 

from the main cordillera. The highest mountain in this group has an 
elevation of 1,110 meters. 

From a geological standpoint the Pico de Loro Mountains, highest 
elevation 678 meters, situated south of the entrance. to Manila Bay, 
helong to the western cordillera. Corregidor and Caballo Islands, 
together v;^ith some small islands and rocks, constitute the evidence of 
a submerged connection. The silhouette of Corregidor Island, which 
is just ^dsible, and the mountains which lie to the north and south on 
the mainland as seen from a distance of about 50 kilometers across 
Manila Bay, are familiar to all who have admired the charm of the 
sunset view from Manila. 

Central plain, — This region is low lying and but little diversified. 
Mount Arayat, just north of the area here discussed, rises as a solitary 
peak to a height of 1,040 meters. On clear days it can be seen from 
Manila. It is situated about equidistant from the western and eastern 
Cordilleras. 

The portion of the central plain included in the geologic map is for 
the greater part occupied by a network of river mouths and esteros. 
These channels constitute the delta of the Pampanga River which has 
its headwaters in the northeastern part of the plains region. The 
Pasig and Orani Rivers of this drainage system are navigable for boats 
of shallow draught which ascend to Guagua and beyond that point to 
Arayat. Near the bay the numerous islands of the delta are covered 
with mangroves and the low nipa palm. 

The area of water-laid tuff lying to the east of the alluvial portion of 
the central plain is appropriately included in the central valley region, 
although it is related to the southwestern volcanic region. This country, 
which is a little higher than the alluvial plain, is but slightly diversified 
excepting near the stream valleys which cross it. It rises gradually to 
the eastward, where it is limited by the foothills of the eastern cordillera. 

Eastern cordillera, — In so far as known the mountain structure of 
this area is with the trend of the cordillera. It is not now possible to 
divide it into minor ranges, and the drainage which is as yet imperfectly 
mapped, does not suggest any important longitudinal valleys. Seen in 
a broad way it presents two physiographic provinces which for con- 
venience may be designated as the northern and southern, the division 
occurring opposite the eastern lobe of Bay Lake. 

The northern division of the cordillera contains a number of con- 
spicuous mountains and subordinate ridges which show alignments 
suggesting ranges. The higher mountains are probably in most cases 
composed of eruptive rocks. The region is not well explored, but within 
the area of the geologic map accompanying this report there are indicated 
Mount Driod, elevation 1,185 meters; Mount Angilo, elevation 1,307 
meters; Mount Batay, elevation 1,408 meters; Mount Banay, elevation 
1,345 meters, and Mount Kanumay, elevation 1,049 meters. 



62 ADAMS. 

The southern part of the eastern cordillera is lower and has a 
peneplaned appearance, and in the region east of the extinct volcano 
Banahao, which is on the border of the volcanic region, the cordillera 
is partially covered with tuffs. There are no conspicuous peaks and the 
higher elevations, in so far as known, fall between 400 and 500 meters. 
To the south of the geologic map the cordillera continues into Tayabas 
Peninsula. 

Lohoo Mountains. — These mountains lie in an irregular peninsular 
area on the south coast of Luzon, and are bordered on the north by the 
plains of the southwestern volcanic region. They are distinguished as 
a separate region because of their discordant trend and the presence in 
them of diorites and tertiary formations which are not represented in 
the adjacent southwestern volcanic region. They appear to form two 
approximately parallel ranges extending in a northeast-southwest (N. 
60° E.) direction and separated by the valley of the Eosario Eiver. 
The northwestern range begins in ISTatoccot Point near which it reaches 
elevations of about 300 meters. It reaches the altitude 987 meters in 
its highest peak, and then dies in the plains. The southeast range 
extends inland from Malabugo Point. Mount Loboo near the coast has 
an elevation of 946 meters. To the northeast the range lowers to eleva- 
tions of about 400 meters and then descends to the plains. 

Southwestern volcanic region, — The larger part of this region consists 
of plains of volcanic tuff sloping toward the sea. Bay Lake, and Manila 
Bay. Deposits of tuff and volcanic breccias and agglomerates are found 
on the east border overlying and obscuring the structure of the eastern 
cordillera. Within the region of the tuff, there are conspicuous moun- 
tains of eruptive rocks, volcanic peaks, the active volcano Taal, a number 
of small crater lakes, and some extinct cinder cones. The highest moun- 
tain is the extinct volcanic cone Banahao, having an elevation of 2,151: 
meters and containing a deep crater which drains to the southwest. The 
high conical peak Banahao de Lucban is on its northeastern flank and 
on its southwestern the small peak Masalacot, and some lesser conspicuous 
hills. Mount San Cristobal lies to the northwest of Banahao and is a 
high extinct volcanic cone containing three small lakes in its crater. 
Banahao and San Cristobal are most impressive as seen from Tayabas 
Bay when approaching Lucena from the west. They can sometimes be 
seen from Manila, a favorable view point being the Ayala Bridge. Mount 
Maquiling is next in importance with an elevation of 1,091 meters: It 
is usually visible from Manila Harbor or the bay front and is conspicuous 
from the steamers which cross Bay Lake, being situated close to its 
southern shore. Its summit has several peaks. When seen from Santo 
Tomas, which lies to the southwest, the peaks define somewhat indistinctly 
the broken walls of a crater. On the slope of Maquiling there are a 
number of small solfataras and at its northern base there is an area of 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 63 

hot springs. At Los Banos these are utilized by the military sanitarium 
and private institutions. There are thus three mountains, Banahao, 
San Cristobal, and Maquiling, retaining evidence of former volcanic 
activity. 

To the east of Maquiling and to the northwest of San Cristobal there 
is an important group of lower mountains lying between Calauan and 
ISTagcarlang. 

The next important mountain is Malarayat-Sosoncambing, having an 
elevation of 1,005 meters. It is a rough, ridge-like mass extending from 
north to south, and lying to the south of Maquiling. Mount Macolod 
on the southeast shore of Taal Lake has an elevation of 958 meters. 
A striking feature of this mountain as seen from the north or south is a . 
high precipice facing Lake Taal. 

Mount Gonzales, to the north of Lake Taal, with an elevation of 749 
meters, has rugged slopes and spurs to the east and northeast. To the 
west it blends with a high ridge of volcanic tuff which runs in a curve to 
the southwest, joining Mount Batulao. This mountain has an elevation 
of 807 meters and its slopes to the west and south form an area of broken 
country. A group of peaks to the northwest of Mount Batulao forms a 
small mountain mass, the highest point of which is Mount Carilao, eleva- 
tion 636 meters. Looking south and a little west from Manila, Mount 
Gonzales is visible, the ridge running from it to Batulao and the tuff 
plains sloping toward Manila Bay. Mount Caralao appears on the 
western limit of this sloping plain in the gap between Mount Batulao 
and the Pico de Loro Mountains. 

This concludes the catalogue of the principal mountains which are 
h^re classified as within the tuff plains. They are neo-volcanic and taken 
as a whole exhibit no well-defined trend. The watershed between the 
streams which flow northward and those which flow southward in the 
tuff plains passes through Banahao, Cristobal, Gonzales, and Batulao. 
The divide does not pass through mountain peaks between Cristobal and 
Gonzales, but takes an irregular course through the plains. On the 
south the streams flow into Balayan, Batangas, or Tayabas Bays. The 
division into these three groups is determined by the mountainous 
peninsular areas separating the bays. The nortward-flowing streams 
enter Manila Bay or Bay Lake. The division into these two groups is 
marked by a ridge presenting an escarpment to the east and extending 
from Mount Gonzales northward to where it is cut by the Pasig River 
which drains Bay Lake, and beyond this continuing to the east of the 
Mariquina River until it approaches the foothills of the eastern cordillera. 
The reasons for these various divisions of the drainage will be made 
more clear when the deposition of the tuff, the elevation of the region 
and the origin of Bay Lake are discussed. 

The remaining mountains of this region are in peninsular areas. The 



64 ADAMS. 

Pico de Loro Mountains on the northwestern part of the region, as 
has already been stated, are geologically a part of the western cordillera. 
In the southwestern part of the region there is an area of high country 
which occupies what is here called the Santiago Peninsula. To the 
north it contains an eminence termed Mount Nasugbu and on its eastern 
border there is a short serrate range presenting an escarpment to the 
eastward. Mount San Pedrino on this range has an elevation of 362 
meters. 

Calumpan Peninsula between Balayan and Batangas Bay is a moun- 
tainous area with its highest peaks rising to elevations of from 320 to 
562 meters. 

The relation of these peninsular areas to the neo-volcanic mountains 
already described is not clear. It appears probable that Santiago 
Peninsula has a core of igneous rocks. Calumpan Peninsula and the 
Loboo Mountain region are separated from Maricaban and Verde Islands 
by rather narrow channels and are not far distant from the high moun- 
tainous island of Mindoro. Moreover, all three of the peninsular areas 
appear to have been made parts of Luzon by the extension of the tuff 
deposits which were not sufficient to fill the intervening bays. 

Taal Volcano, surrounded by a number of small extinct parasitic 
cones, is near the center of an island in Lake Taal. The highest point 
of the island is on the southwest rim of the crater which reaches an 
elevation of 320 meters. The floor of the crater is at about the same 
level as the water of the lake and but little above sea level. Mention 
has already been made of the high ridge connecting Mount Gonzales and 
Mount Batulao. Its slope to the north is gradual, but on the south it 
piresents an escarpment along its eastern part which faces Taal Lake, 
and further to the west the country to the south of its crest is broken 
and dissected by deep valleys in the part adjacent to Mount Batulao. 
The country around the southern border of the lake lies at elevations 
varying from 200 to 300 meters near the lakers shores, which in many 
places are steep and even precipitous. Mount Macolod when seen from 
a distance appears to have a precipitous face rising from the shore of 
Taal Lake. On nearer examination this is not so striking a feature. 

The area lying between M^aquiling and Malarayat Mountains on the 
west and Cristobal and Banahao on the east, may be called the crater lake 
country. Within it, distributed in an irregular way, there are twelve of 
these lakes, varying from a half to more than one kilometer in diameter. 
They lie in the tuff plains, their rims rising but a few meters above the 
general level of the country. There is but little evidence of them until 
one stands upon their borders. The coconut groves and other vegetation 
disguise their presence so that passers by on a road within a minute's 
walk of one of these lakes may not know of its existence. The lake to 
the north of San Pablo may be reached by walking to the end of one of 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 65 

the streets of the town. Some of the lakes are probably very deep, 
while others are shallow and show evidence of filling up, and in one 
case a part of the bed is cultivated. The walls are usually steep and 
the level of some of the lakes is as much 30 or 40 meters below the rim. 
Outside this preeminently crater-lake country there is a well-known 
crater lake on a peninsula just to the west of Los Baiios. A small one 
to the southwest of Nasugbu, probably lies in the remnants of a once 
well-defined crater. 

A few cinder cones are shown on the geologic map. There is one to 
the northeast of Batangas, another is situated just north of New Rosario, 
and a third lies to the south of Lipa. These small hills have a conical 
appearance when seen from a distance, but upon close examination are 
found to have a crescentic plan, or a central depression indicating the 
position of the crater. Tuluc Hill, between Santo Tomas and Calamba, 
near Bigaa and to the west of the road, is probably a cinder cone, but 
shows no crater. La Mesa Hill, south of Calamba, is a cinder cone with 
a slight depression indicating a crater. Around it there is a crescentic 
outer crater. Similarly, Pansol Hill and Mayzondon Hill, near Los 
Banos, are probably remnants of cinder cones. A careful survey will 
probably disclose others. 

The stream valleys in the area of the tuff plains are consequent and 
nearly all flow in deep, narrow, gorge-like valleys for the greater part 
of their courses, so that the best routes of travel lie between the streams. 
To the north of Mount Banahao the country adjacent to the eastern 
Cordillera is an upland. Nearly all the streams which flow from this 
area have picturesque falls or gorges. Pagsanjan Gorge and the falls 
near Pagsanjan, as well as Botocan Falls between Majayjay and Luisiana, 
are noted for their scenic beauty and are visited by many tourists. In- 
deed, Taal Volcano and Lake, the hot springs at Los Banos, the crater- 
lake country and the falls and gorges just mentioned, combined with 
the imposing views of volcanic mountains, make this region one of the 
most inviting to tourists in the Philippines. The extension of the Ba- • 
tangas railway lines and the constructing of more good roads will 
certainly make it one of the most interesting scenic districts of the Is- 
lands. 

GEOLOGY. 
WESTERN CORDILLERA. 

Geologic mapping. — The only geologic mapping within the region of 
the western or Zambales cordillera which has previously been published 
is by Von Drasche, Plate I. He quotes Eoth^s statements in regard 
to the rocks of the Natib, Mariveles and Pico de Loro Mountain masses 
and says that, accepting this, he colored the corresponding part of his 
map as dolerite. He did not visit the area. Further north, and just 



66 ADAMS. 

beyond the area discussed in this report, he ascended the foothills of th6 
Cordillera which form a ridge to the west of Porac. He colored the 
area corresponding to the axis of this ridge as dolerite and surrounding 
it he showed a border of doleritic tuif . The remaining descriptions of 
the western cordillera by Von Drasche are accompanied by only a route 
map. ' Becker has recently studied some rocks from the region which 
Eoth classed as dolerites and, with the exception of a dike of dacite on 
Corregidor Island, has pronounced them andesites. In this report the 
igneous rocks of the portion of the western cordillera which is included 
in the geologic map are classed as "principally andesites/^ The larger 
alluvial areas ai^e shown and the occurrence of marine conglomerates 
lying on the flanks of the andesites is partially indicated. 

Besides descriptions of the occurrences of various types of igneous 
rocks, there are, in the literature concerning the northern part of the 
western cordillera, references to older sedimentaries (slates), limestones, 
and a tuffaceous foraminiferal marl. These formations will now be 
discussed under separate headings. 

Older sedimentaries. — Abella says, in describing the geology of the 
vicinity of the hot springs at Canan which are situated 3^ kilometers 
west of O^Donnell and to the north of the limits of the geologic map 
accompanying this report: "The land in which all these springs appear 
is of an alluvial nature, but of little thickness in the vicinity of Canan, 
since in the bed of the Malibi and the lower slopes of Cosipen and 
Marangla there appear strata of compact clayey slates with an ancient 
appearance striking ENE. and WSW. with dips of 25° NNW. This 
slate formation is cut by dikes of beautiful sanidine-trachyte with a 
porphyritic texture which appear on one side in slopes of the Hills 
Marangla and Cosilen and on the other between the town and the 
Capatian River forming the volcanic line of the Hills Dayagdag, Taogan, 
and Patlin.'^ 

This is the only mention of a locality of older sedimentaries in the 
• Zambales Mountains which has been noted. However, it should be 
remembered that Yon Drasche says that the Porac River brings down 
hard, flinty slates. 

Igneous rocks.^ — The specimens of igneous rocks which were collected 
from the western cordillera during the progress of field work are 
principally andesites. There are some basalts and in one locality a 

^ The writer is not a competent petrographer but in the course of his field work 
collected hand specimens of the various rocks. The microscopic determinations 
have been made by Dr. W. D. Smith and Mr. H. G. Ferguson, whose assistance 
and cooperation is gratefully acknowledged. In this number of this Journal a 
petrographic description of some of these rocks will appear by Prof. J. P. 
Iddings, who has kindly advised as to the petrographic part of this report. 



GEOLOGIC RECONNAISSANCE OP SOUTHWESTERN LUZON. 67 

dacite. Besides these extensive rocks there are gabbros and peridotites 
and some metadiorites. 

The southern part of the cordillera in Bataan Province was mapped as 
dolerite by Von Drasche following the statements of Both. Becker, 
who examined rocks found near Mariveles, classified them as andesites. 
He says that they vary from the basaltic to the trachytic type, but they 
are all essentially labradorite rocks without ophitic structure and that 
for the most part they are pyroxene-andesites containing augite and a 
dichroic^ rhombic pyroxene, but one gray rock quarried (at Sisiman) 
for use in Manila is hornblendic. He says that he is rather explicit 
about the rocks of this locality because Both calls the rocks of this 
region dolerites. Further on he remarks that the rocks of the Semper 
collections from Mariveles correspond with his own in most respects, 
being labradorite rocks more or less glassy without ophitic structure, 
and that portions of the Mariveles rocks examined by Oebbecke were 
found to carry hornblende in addition to augite, and some grains of a 
mineral, which is probably olivine, were detected. The Pico de Loro 
Mountains south of the entrance to Manila Bay were classed by Both as 
doleritic. The specimens obtained from this area are andesites* in 
every respect similar to those of the Mariveles coast. 

The dacite which was discovered by Becker on Corregidor Island was 
described by him. He says in part : "It is blindingly white and shows 
to the naked eye besides feldspar only some quartzes and minute opaque 
mica scales. . . . The bulk of the rock is made up of feldspar micro- 
lites with just a sprinkling of magnetite and a little apatite, . . /' 

Yon Drasche crossed the cordillera to the north of the area here under 
discussion. In describing the structure along his route from O^Donnel 
to Iba, he says : "It appears that the sierra consists principally of dioritic 
and peculiar diabase gabbro rocks which often exhibit bedded structure. 
These rocks are in intimate relation with ordinary gabbros and serpentines. 
On the eastern slope of the southern half of the sierra one finds these 
rocks overlaid by a thick, trachytic tuff formation which incloses numerous 
fragments of trachyte. This tuff can be traced up to the watershed 
(pass) at a height of 3,000 feet and to the east it is related to the plains 
of Pampanga (central valley), the floor of which consists principally 
of the weathered products derived from it. The crystalline rocks must 
be broken through at numerous points by trachyte, since such rocks 
are found in great numbers among the bowlders derived from the sierra/' 

He describes especially what he called a sanidine hornblende-trachyte 
found near Porac, and Abella the occurrence of a similar "sanidine 
trachyte" with porphyritic texture which he found cutting older slates 
near O^Donnel. Smith has shown this rock to contain chiefly plagioclase 
feldspars and to form Mount Pinatubo. At Olongapo there is a small 



68 ADAMS. 

hill of it near the naval station which is quarried for use as crushed 
stone. It is found as bowlders in many streams and it undoubtedly 
constitutes a large part of the cordillera in the region of the Pinatubo 
group of mountains. It is possible to identify Von Drasche^s "trachytes" 
by means of the distribution and habit of this rock. Usually it has a 
coarse porphyritic texture, contains conspicuous glassy feldspars, and 
is very friable. 

The diabase gabbro rocks mentioned by Yon Drasche are represented 
in the collection at hand. Some gabbros, peridotites pyroxenites and 
serpentines were found near Subig in the Cinco Picos Mountains. 

Basaltic rocks were found to the west of Floridablanca at the foot of 
the cordillera, in the vicinity of Subig, and in Mariveles Mountain, but 
they are not conspicuous. 

Hochstetter, in his map of the volcanoes of Luzon, indicates Corregidor 
Island, Mount Mariveles, Mount Batilao or Natib and Mount Pico 
de Loro as extinct volcanoes. As will be explained in this chapter, 
Corregidor consists largely of marine conglomerates. Mount Mariveles 
has somewhat the appearance of a volcanic cone and the form of its 
summit suggests a crater with the northern part of its rim broken down. 
The quarry at Sisiman on the coast at the south base of the mountain 
shows columnar structure, and the hill in which the quarry is located 
may be a small volcanic neck or stock. There are no solfataras in the 
western cordillera and volcanic activity has long since ceased. The 
younger igneous rocks are eruptives and at certain centers, as for example 
Mount Mariveles and Mount Pinatubo there probably at one time were 
volcanoes which gave origin to the tuffs found on the flanks of the range, 
but it certainly is not correct to class these eruptive centers as extinct 
volcanoes along with those which are found in the southeastern and 
southwestern volcanic regions of Luzon. 

Limestones. — Eoth cites Gallery as reporting the occurrence at a dis- 
tance of 4 leagues west of Sual, of a belt of coarse limestone and travertine 
having a width of 2 leagues, resting horizontally and containing decapod 
crustaceans. This locality which is near the northern termination of 
the western cordillera and on its eastern base not far from the Gulf of 
Lingayen, is too distant from the area discussed in this report to warrant 
examination at this time. It is the only occurrence of a limestone 
formation which is at all definitely reported from this cordillera. The 
fact of its being horizontal suggests that it is probably younger than the 
cordillera and, without wishing to prejudice future observers, it is sug- 
gested that it may be the equivalent of the tufaceous foraminiferal marls 
described by Von Drasche as occurring near Santa Cruz on the west side 
of the same mountains. In passing it may be well to note that Von 
Drasche says that besides the marls he found no sedimentary formations, 
but the cura of Porac told him that limestones were found in the cor- 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 69 

dillera near that place. The writer, after considerable field work, has 
concluded that the euro's statement may be dismissed as having no value 
in the literature. 

Smith, who made a hurried trip through Subig, on his way to ascend 
Mount Pinatubo, states in his report that he judged from the appearance 
of the country that there are limestones and shales in the region sur- 
rounding the bay. The writer found none and was unable to learn of 
any through inquiry. 

Finally it should be mentioned that Eoth, without any seeming basis, 
and without any subsequent substantiation ' unless it be the statement 
by the cura reported by Von Drasche, says, in discussing the occurrence 
of limestones in Luzon, that probably the abrupt peaks in the region 
which corresponds to the position of the Pinatubo group of mountains 
are composed of the same limestones which are found near Antipolo and 
Bosoboso in the eastern cordillera. This statement is without value 
since the peaks are now known to be formed of eruptives. 

Marine conglomerates. — The only . reference to a locality of marine 
conglomerates on the flanks of the western cordillera is in the report on 
the Sisiman quarry by Ickis. He says, "along -the path north of the 
blacksmith shop and directly below Mr. Nelsoh^s home, occurs a 10-fpot 
bank having the appearance of a conglomerate, while further north on 
the same path the bank is very soft resembling tuff. Both these occur- 
rences may result from weathering and washing down of the andesite." 
Ickis did not recognize the significance of this conglomerate as he 
probably would have had he seen more typical exposures. It is found 
east of Sisiman Bay and also on the point west of the entrance to Ma- 
riveles Bay, where in a cove it forms a sheer bluff 30 meters high. The 
writer first studied the conglomerate on Corregidor Island, which locality 
will now be discussed. 

An erroneous idea in regard to the character of Corregidor Island is 
presented by several writers, none of whom, however, has made a careful 
examination of the locality. It seems to have had its inception in the 
statement by Von Kotzebue that it contains a crater. This idea, which 
has even become popular, is most fully elaborated by Von Drasche, who 



"The entrance to the bay has a width of 10 sea miles but is narrowed by two 
islands. One of them, Corregidor, is nearly three and a half sea naijes long and 
is composed of lava beds. Von Kotzebue, according to J. Roth, mentions a crater 
upon it. Although I likewise did not visit this island, nevertheless, I had 
opportunity to pass both to the north and south of it and noticed only slightly 
dipping lava beds. To the south of it lies the small island, Pulo Ciaballo." In 
his paper he has a text figure which shows both islands with the soundings taken 
from the chart of the Bay of Manila by Claudis Montero. 

"According to this sketch it appears that the two islands were once joined to the 
east, but now they are separated by a depth of 69 meters. Also the soundings 
suggest that the two islands once formed a circular mountain with a deep crater 



70 ADAMS. 

within. A close examination of the direction of dip of the lava beds would give 
definite conclusions in regard to this." 

To the west of Corregidor there is the small volcanic island La Monja and to 
the south the rock El Fraile. These also show clearly that they are formed of 
lava." 

Becker refers to the statements of Von Drasche which suggest that 
Corregidor and Caballo Islands are parts of the rim of a large crater and 
says that this also appeared probable to him during his visit to these 
islands. 

The only record of an actual examination of any part of Corregidor 
appears in the writings of Becker, who says: "I found an interesting 
dacite on the Island of Corregidor, at the entrance to Manila Bay. It 
forms a wide irregular dike crossing the island from north to south 
just east of the little settlement of San Jose, and terminates in a bold 
white cliff visible from the south passage or Boca Grande." He described 
in detail the appearance and microscopic structure of the rock. 

This is the only exposure of igneous rock which was seen on the island. 
Becker says the dacite is a dike but he does not state what rocks it cuts. 
With the exception of the bluff on the south side of the island it is 
surrounded and partially overlaid by marine conglomerates with a 
variable matrix which grades into thick beds of finer sediments. The 
conglomerate is exposed in vertical bluffs around the western part of 
the island and on the eastern part it is weathered and shows stratification 
and slight folding and some faulting. On the main part of the island 
the conglomerate was seen well exposed in the excavations made for 
fortifications up to an altitude of about 150 meters. The highest point 
of the island is about 200 meters above the sea. There are no good 
exposures on this dome-shaped portion, but there are numerous bowlders 
lying on the surface and imbedded in the surface clay. These are 
andesites and it is probable that there is a mass of andesitic rock which 
forms the core of the higher, western part of the island and the bowlders 
in the conglomerate have been derived from it. There are no lava beds 
on the island, but the matrix of the conglomerate and the beds of finer 
sediments contain tuff and volcanic ash evidently from distant sources. 
The eastern part, or tail of the island, as it is called, has an average 
elevation of about 75 meters and in so far as seen consists of conglomerate 
beds. The present position of the marine conglomerates indicates an 
elevation of Corregidor Island of about 200 meters. They are the 
equivalent of those found on the mainland near Sisiman and Mariveles. 

Exposures of conglomerate w^ere seen up to an elevation of 100 meters 
at Olongapo on the ridge to the north of the Calacan Eiver and to the 
east of the rifle range, and there is a bluff of conglomerate facing the 
lowland northeast of Subig. 

By far the most important area of this formation is to be found on 
the north and west flanks of the Pico de Loro Mountains to the south of 



GEOLOGIC RECONNAISSANCE OP SOUTHWESTERN LUZON. . 71 

the entrance to Manila Bay. The lower slopes of the mountains descended 
to the coast gradually and at the shore are cut off abruptly in sea cliffs. 
In these cliffs the marine conglomerates are well exposed. West of Ter- 
nate they are in many respects similar to the beds on Corregidor Island. 
They were also seen in the cliffs of Carabao and other islands near the 
shore and in the jagged coast line south, to the vicinity of Nasugbu. 
In this part of the coast, which is exposed toward the China Sea, they 
have yielded more readily to erosion, and deep inlets and coves have 
been formed. The conglomerates are found up to an elevation of about 
200 meters north of Nasugbu. The bowlders in the conglomerate are 
mostly andesitic. The finer materials are in part derived from erosion 
of igneous rocks, but to a considerable extent they are tufaceous. Some 
of the tuf aceous material may have come from the (^nudation of deposits 
on the mountains of the western cordillera, but it is probable that much 
of it is the equivalent of the great tuff area found in the southwestern 
volcanic region. This is suggested by the fact that the Pico de Loro 
Mountains on their eastern flank are partially overlaid by water-laid 
tuffs. Near Ternate and Naic the conglomerates apparently grade into 
the deposits of water-laid tuffs which have a great extent in the adjacent 
plains. 

Tufaceous foraminiferal marls. — A tufaceous foraminiferal marl, 
which was discovered and described by Von Drasche, occurs on the 
western base of the sierra along the coast from Palanag to Santa Cruz 
and perhaps further north. This author reports it as extending up to 
' an elevation of 400 feet. Felix Karrer determined the fossils from this 
formation and assigned it to the younger Miocene. This formation may 
correspond to the marine conglomerates which are found on the flanks 
of the southern portion of the cordillera. 

Geologic history, — The early history of the western cordillera and the 
succession of the igneous rocks which constitute its principal area is not 
yet plain. The younger sedimentaries on its flanks, however, give a key 
to some of the later changes which have taken place. Von Drasche, who 
crossed the northern part of the range, says: "There may have been a 
time when the sierra Zambales (western cordillera) stood in the sea in a 
form similar to that which Paragua Island (Palawan) now has. On 
the west of the island the foraminiferal marl was built up from the 
volcanic detritus under the action of the sea, while on the east it was 
covered by a heavy mass of tuff mixed with bowlders of volcanic rock. 
Through continued uplift the great central plain of Luzon finally became 
dry land and apparently it is even yet rising.^^ 

The relations of the marine conglomerates described in this report 
furnish additional proofs of the supposition of Von Drasche, which was 
based upon the evidence of the foraminiferal marls, and indicate an 
emergence of the cordillera amounting to as much as 200 meters. In 



72 



ADAMS. 



the accompanying sketch map, fig. 2, the author has attempted a recon- 
struction of the land area of southwestern Luzon as it was before the 
emergence of the sediments found on the flank of the western cordillera. 
The Pico de Loro Mountains were probably an island as were also those 
of the Mariveles and Cinco Picos groups. The Nat^b Mountains formed 
a peninsula from the mainland which included the Pinatubo group. 




«^-^ 



Fig. 2. — Sketch Map Showing Probable Land Areas Before the Emergence of 
THE Central, Plain and the Plains of the Southwestern Volcanic Region. 

Corregidor Island was low lying and probably, as erosion progressed, it 
was reduced to a shoal. The area corresponding to the central plain of 
Luzon was a broad strait. 



CENTRAL PLAIN REGION. 



The portion of this region included within the area described in this 
report is occupied principally by an alluvial and littoral formation. To 
the west it borders on the igneous formations of the western cordillera; 




GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 73 

to the east, the alluvium overlies a water-laid tuff formation which in 
turn extends to the foot of the eastern cordillera. 

Inasmuch as this region consists very largely of low-lying land, there 
are but few exposures in which the formations may be studied. However, 
fortunately, a large number of wells have been drilled and the records 
of these throw some light on the stratification' of the alluvial deposits 
and their relation to the underlying marine sediments which are not ex- 
posed and the tuff which forms a belt on the east side of the region and 
extends under the alluvium. No wells have been drilled in the western 
part of the area and there the relations are not so well understood. 

Water-laid tuff. — This formation is a continuation of the tuff which 
has a wide extension in the southeastern volcanic region. It is usually 
clearly stratified and exhibits beds of variable thickness. In places it 
grades into clayey, somewhat shaley beds and it oc- 
casionally contains a conglomeratic phase, especially 
near the footliills of the eastern cordillera. It is ^ 
probable that a large part of the tuff deposits was 
thrown out by the volcanoes of the southwestern 
region, but certainly some sediments must have been 
derived from the adjacent cordillera. 

The records of the wells which have been drilled 
in the tuff show occasional beds of marine sands ^^%qlaiid^^^)^ ^ fo^und 
and some strata which are composed of waterworn in a railway cut in 
gravels and fine pebbles. Occasionally a log of 7rell'""^of \h^ ° n^ge 
wood has been encountered in drilling, and plant between Bay Lake 
remains, fish teeth (fig. 3), and one mammalian ^"^ ^^""^ ^^^- ' 
tooth have been found in the beds. The presence of the plant remains 
has been recorded by many observers. The greatest depth at which a 
log of wood has been fotind is in the Alabang wells south- 
east of Manila, where on was cut by the drill at a depth 
of between 130 and 132 meters. The mammalian tootli 
(fig. 4) was obtained from the Pasig well at a depth 
somewhere between 81 and 85 meters. 

No tuff formation is exposed on the western border pio. 4.— Mamma- 

^ ,, T J • i-u- l*a^ tooth (cf. 

of the central plains withm the area discussed m this antelopes of the 
report. It will be remembered that the marine conglom- siwaiik piio- 

^ -.,, cene of India) 

crates on the flanks of the western cordillera are cor- ^^und at a 
related with the tuff deposits of the southeastern volcanic depth of be- 
region, but these conglomerates have not yet been seen gg meters in 
adjacent to the central plain. Von Drasche saw, at ^^^^^^^^^ * ""^^^ 
Porac to the north of Floridablanca, a tuff formation 
in an excavation that had been made for laying the foundations of 
a church. It there consists of layers of sand with fragments of rock, 
similar to those commonly presented in the western cordillera, and clay 
beds interstratified. Von Brasche was of the opinion that the fragments 

93217 2 




74 ADAMS. 

of stone should be classed as volcanic bombs. Near Porac he discovered 
a dolerite tuff formation, which he described as extending up onto the 
flanks of the cordillera to an elevation of 1,000 meters. He doubted if 
the whole formation had been deposited in water. The part of this tuff 
formation adjacent to the plains near Camp Stotsenberg is probably 
>v^ater laid. Both Porac and Camp Stotsenberg lie well to the north of 
the area which is mapped in this report, but the formation obtained at 
these localities proves that the tuff formation extends under the alluvium 
and it is probable that it will be encountered in drilling wells near the 
western border of the plains. The exposures of the tuff near Camp 
Stotsenberg differ so materially in their general appearance from the 
tuffs of the eastern border of the central plains that they can not be 
correlated on lithologic grounds. 

Alluvium. — The part of the central plain included in the geologic 
map is very largely occupied by the delta of the Pampanga Kiver, which 
forms a network of channels dividing the delta into many islands. 
These island areas are but little above sea level and many of them are 
partially overflowed at high tide; those lying lower are occupied by 
groves of mangroves and nipa palms. West of the delta the alluvium 
rises toward the mountains and has the form of an aggraded plain. The 
part of the plain which receives drainage from the western cordillera 
may be distinguished by the presence of the plagioclase sand and frag- 
ments of andesitic rocks which give rise to the sandy soil. The eastern 
border of the delta system is limited by contact with the tuff formation, 
and the sediments which reach it from the eastern cordillera are car- 
ried by streams which flow in deep, well-defined channels. The larger 
streams which head in the mountains carry sand and gravel. There is a 
deposit of gi-avel at Caloocan which probably represents a former delta 
deposit of the Tinajeros Elver, which now flows to the north of the towTi 
at a lower level. This bed of gravel, as shown by the records of wells at 
Caloocan, is about 30 meters thick. The sediment brought by the Pam- 
panga River to the head of the delta is nearly devoid of gravel. 

While the alluvial deposit covers a large part of the central plains as 
a thin veneer, underneath it there is a series of beds which are marine, 
or at least deposited in brackish water. These beds are encountered in 
drilling wells and may be distinguished by the presence of numerous 
shells and marine silts and muds. At Tarlac, which lies near the eastern 
border of the plain and a little more than halfway from Manila to 
Dagupan, shell beds are found at a depth of about 4 meters. Details of 
their occurrence are given by Centeno, who states that shafts were sunk 
to the beds in order to obtain the shells for burning lime. This oc- 
currence, together with the records of the wells, would seem to prove that 
marine sediments underlie a large part of the plain. It is probable 
that a strait once extended from Manila Bay northward to Lingayen 
Gulf and that the gradual elevation of the Island of Luzon transferred 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 75 

this area into low-lying land and it has been gradually aggraded and 
extended by the sediments brought by the river system. 

The southern end of the Candaba swamp which has a larger extent 
beyond the limits of the geologic map, lies on the eastern border of the 
alluvial area, between Baliuag and Apalit. A part of this swamp is 
included within the alluvial area of the geologic map. During the 
rainy season it is partially overflowed, but during the dry months a 
large portion of it is cultivated. There has been some talk of attempting 
to reclaim it to cultivation by means of drainage canals. 

Various opinions are current in regard to the origin of this swamp; 
one is that it represents a remnant of the retreating sea which has been 
freshened, and, in support of this»view, it is stated that it contains a 
brackish water fauna. However, since it receives a large amount of 
fresh water during the rainy season, and has outlets, there is little reason 
to believe that it contains any salts which remain as a result of the 
impoundiilg of an arm of the sea. A second theory in regard to this 
swamp is that it represents an area of subsidence. This is not very 
fully discussed in any writings, but is somewhat substantiated by Centeno's 
report on the earthquake of 1880, in which he records that a great many 
cracks opened in Nueva ficija Province along a zone which, if continued 
southward, would pass near the Candaba swamp. A simple explanation 
is that the Candaba swamp represents an area which lying between the 
main channel of the Pampanga Eiver and the Quiangan Kiver, has failed 
to receive sufficient sediments to build it up as rapidly as the remaining 
portion of the delta. 

EASTERN CORDILLERA. 

Some of the earlier descriptions of the formations of this region are 
faulty, and certain of the conclusions reached are based' upon reports 
that have since proved untrustworthy and conjectures which were sup- 
ported by no reliable information. This is especially true in regard to 
the so-called older crystallines and slates and the possibility of the 
presence of the Carboniferous. 

Schists. — On Yon Drasche^s map two areas of older crystallines are 
shown in the eastern cordillera. The color representing them bears the 
sub-legend "Gneiss, Chlorite-homblende schist, etc." The northern area, 
lying to the east of Angat, was not visited by Yon Drasche, and it is 
probable, as Becker has suggested, that this area was mapped on the 
strength of Itier^s statement which is quoted by Yon Drasche as follows : 
"In Angat at the base of the foothills of the cordillera of Luzon, there 
exists no evidence of volcanic products, and the rocks carried by the 
river are diorite amygdaloid, spilite amygdaloid^ diabase, epidote, dolomite 
and porphyry . . . ." Both says that the iron deposits near . Angat 
indicate the presence of crystalline slates, but his course of reasoning is 
not clear. It is possible that he was influenced by the fact that iron 
deposits are found in the Camarines associated with older crysitallines. 



76 ADAMS. 

' McCaskey, who made a reconnaissance of the region of the iron mines 
near Angat, lays special stress on the fact that his observations justify 
Von Drasche's mapping of the area as older crystallines, but he reported 
finding fine-grained diorite, a granulite (quartz and feldspar with horn- 
blende or mica), a gabbro, and a rose-red trachyte. He makes no 
mention of gneisses or schists, and accordingly it is questionable if any 
are to be found. 

The southern area of schists mapped by Von Drasche lies to the 
southwest of Atimonan, and was seen by him in crossing from Laguimanoc 
to Atimonan. He says, "On the eastern slope one finds finally a green 
schistose chloritic talcose, but very decomposed rock, that in general 
has a steep dip to the eastward, while the tuff beds for the most part 
dip to the southward. ^^ 

In crossing the eastern cordillera from Lucena via Pagbalao, to 
Atimonan, a micaceous schist was found about 5 kilometers west of Ati- 
monan. On the accompanying geologic map this outcrop is included 
with Von Drasche's locality in a single area. The significance of these 
schists is not clear, but it is quite possible that they are metamorphosed 
Tertiary sediments. 

Diarites. — Within the area here discussed, the first locality mentioned 
as containing older rocks is the region to the east of Montalban, where 
the dam for the headworks of the water supply of Manila is now being 
constructed in a gorge between two limestone cliffs on the San Mateo 
River. In the literature, this is spoken of as the limestone cave region 
of San Mateo, although it is distant from the latter place. Becker cites 
Meyer as observing that the limestone in which the caves are found 
rests upon diorite. Von Drasche, who visited the locality later, did not 
describe such an occurrence, but he says : "The bowlders of this river 
consist for the most part of older syenitic rocks, diorite, etc., and a 
little trachyte.^^ ^e described the relations of the limestones a short 
distance to the north (Poray Creek), but made no mention of diorite 
near it. In making the reconnaissance on which this report is based, 
no diorite was found at the locality of the limestone. 

Itier says : "In Angat, at the base of the foothills of the cordillera of 
Luzon, there exist no evidence of volcanic products and the rocks carried 
by the river are diorite, amygdaloid, spilite (amygdaloidal diabase), epi- 
dote, dolomite, porphyry , , ." 

Some of the streams of the eastern cordillera bring down diorite peb- 
bles. Diorites are now known to occur in the region. At the iron 
deposits on the Lanatin River, worked by the Spanish and described later 
in this report, the writer found an area of diorite, a part of which is 
within the drainage basin of the river that passes Montalban. Near the 
are?L of schists west of Atimonan, dioriti<3 rocks are exposed. No doubt 
Qther 3-reas. will, be discovered whep the country is thoroughly explored. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 77 

Becker, in summing up what was known concerning the older rocks 
of the Philippine Archipelago, said that there is no inconsistency in 
Abella^s generalization that the dioritic rocks are the oldest in the 
Philippines if the schists of similar composition are included under the 
head of dioritic rocks. In the reports by Eveland on the Mancayan- 
Suyoc mineral region and the Baguio mineral district, the basal rocks 
are stated to be diorites.; thus concurring with Abella's views concerning 
the central Cordillera of Luzon, on the western slope of which these 
districts are located, far to the. north of the area here under discussion. 
However, there are no grounds for this conclusion. The oldest rocks so 
far as now known are a complex containing gabbros accompanied by 
periodites and pyroxenites besides some diorite and granite. 

Andesites. — In the northern part of the region under discussion, 
andesitic pebbles are very commonly found among the stream gravels, 
and many of the mountains have probably been formed by the eruption 
of andesites. The andesites in Jalajala Peninsula, which the writer has 
seen, were described by Both as dolerites. The northward continuation 
of this ajea was crossed by Ickis, who indicated its borders in his section 
from Infanta to Tanay. Ickis also described an area of andesites further 
to the east near Infanta and separated from the one just mentioned by 
a belt of sedimentaries. On the geologic map accompanying this report 
these areas are indicated approximately, and also some additional ones 
which were seen by the writer in traveling from Antipolo to Santa Inez. 

Basalts. — A greenish, ophitic basalt outcrops in the vicinity of Ango- 
no, and extends northward passing to the east of Taytay, and is exposed 
in some of the railway cuts on the line between Taytay and Antipolo. 
In places the rock has a fine-grained texture and is frequently brecciated 
and in places is amygdaloidal. The outcrops indicate an eruptive mass 
trending a little to the west of north, the basalts and the limestone belt 
further to the east being approximately parallel with it. To the south- 
east of. Angono this rock is overlain by later basaltic tuffs and breccias. 
The best known occurrence of basaltic rocks in the part of the eastern 
Cordillera covered by this report is situated in the peninsular area on 
which the towns of Binangonan and Morong are located and in Talim 
Island, where they have been extensively quarried. Becker has published 
petrographic descriptions of typical basalts, which he collected on Talim 
Island. Hochstetter described Punta Gunong Bajang and Punta del 
Diablo by Binangonan as formed of the most remarkable obsidian columns. 
An examination of these places by the writer failed to verify his state- 
ments, but basalt . showing flow structure was found. Both, evidently 
following Hochstetter to some extent, considered the "doleritic lavas^' of 
Talim Island and those in Jalajala Peninsula, together with the oc- 
currence of "obsidian^^ near Binangonan_, indicating a grand volcanic 
center. Hochstetter's ideas concerning this is expressed in Von Drasche's 



78 ADAMS. 

quotation, which in part is : "all indicates a great volcanic center which 
is lacking, but should lie where now the deep Kinconada Bay is sunk/^ 
The name Kinconada refers to the middle lobe of the lake. Von Drasche 
also says that Hochstetter unites this center of eruption with the Ma- 
quiling and Malarayat-Sosoncambin Mountains in one system of volcanoes 
which lie on a north-south fissure. He also quoted Eoth quite fully 
concerning this area and then remarks that these observations correspond 
very litne with the view that an eruptive center should be sought in the 
Einconada. 

One of the prevalent ideas concerning Bay Lake is that it was formed 
by subsidence and may possibly be a crater or caldera similar to Taal 
Lake. Dana discovered no evidence that the lake corresponds to a 
single crater, and further states that Talim Island is probably a volcanic 
summit, and another small island ("Pulo Bay^^) off Bay consists of the 
lavas of another. It seems that Hochstetter was correct in considering 
that the eruptions occurred along the north-south line, since his field 
observations show that the basalts extend northward to beyond Antipolo, 
forming a rather definite zone bordered by tuffs and agglomerates. The 
falls of Antipolo flow over this basalt. The origin of the lake is discussed 
in a subsequent portion of this report. 

The "trachyte" described by Richthofen as inclosing the limestone 
northeast of Binangonan is a porous basalt often vesicular and frequently 
brecciated. Basaltic flows, breccias and tuffs intermingled in great con- 
fusion are found in the Binangonan-Talim zone. Susun-dalaga, the 
highest peak of Talim Island, exhibits all these phases. Certain rocks 
from this peak approach an andesite in composition. To the east of the 
peak there is a bay in the island. The configuration of the island at this 
point and the position of the islands near by have suggested to some 
observers that the bay marks the position of a crater. No definite evi- 
dence of a crater at this place was seen, but the peak is an important 
point in the eruptive zone. The relations of Maquiling and Malarayat 
Mountains to this zone will be discussed more fully in the part of this 
report which deals with the southwest Volcanic region. 

Some basalts and basaltic agglomerates were found near Tanay and 
Pililla. They are probably related to those of the area just considered. 
Ickis noted a dyke of basalt a short distance west of Tanay. Becker 
described a labradorite basalt which he collected near Paete on the east 
shore of the lake. Basalts occur at several places along the same shore 
of the lake, and it is probable that they extend southward into the tuff- 
covered area. 

Basalts were also seen just east of the landing at Mauban on the 
Pacific coast and pebbles of basalt were, found in a conglomerate on the 
road from Mauban to Lukban. Large bowlders occur sparingly on 'the 
adjacent uplands. Many streams of the eastern cordillera carry some 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 79 

basaltic pebbles and it is probable that there are numerous dykes of it in 
the region. 

Tertiary limestones, sandstones, shales, and coal, — The first mention 
of the occurrence of limestones in the eastern cordillera was by Meyer. 
Eoth cites him as saying that he saw at San Mateo (probably the cave of 
San Mateo and present site of waterworks dam) and near Balete (Mon- 
talban) fine-grained hornblendic slates lying on the limestones of San 
Mateo. Following this is the description by Von Riqhthofen. He 
visited the locality of the San Mateo cave, at which place he saw a 
limestone mass extending far to the north and inclosed between trachytes. 
In a direction nearly south from this place he saw an isolated limestone 
hill surrounded by trachytes, half way on the road from Antipolo to 
Bosoboso. In extension of this line he found a small limestone mass 
which he described as rising in an arrow-like form out of trachyte and 
lying northeast of Binangonan. The limestone is stated to be older than 
the trachytes since pieces of it were found in a trachyte breccia. 

Von Richthofen says concerning the age of the limestone that from 
outward appearance it would be considered as Jura, but that he was so 
fortunate as to find Nwnmulites and therefore referred it to the Eocene. 

Von Drasche found limestones on Poray Creek at which place he saw 
a compact breccia composed of limestone and diabase-aphanite on both 
sides of it. He conjectured that this exposure is a continuation of the 
limestone to the south at the San Mateo cave, which he also examined. 

In the geologic map accompanying this report, the limestones at the 
localities of Poray Creek, the gorge at the waterworks dam site, the 
exposure between Antipolo and Bosoboso and the Binangonan locality 
are included in a single area. 

The next reference to the limestone of this belt is by Smith, who 
found Orbitoides in some samples collected by Ickis from the Binangonan 
locality. Smith visited this locality, collected additional specimens and 
described Orbitoides richthofeni s. n. He believed this to be the fossil 
which Richthofen called Nummulites. He called attention to the fact 
that Becker considered as unsatisfactory the evidence showing the 
Binangonan limestone to be Eocene as classified by Richthofen and that 
there is no reason why it may not be Oligocene or even Miocene. 
Further supporting himself by the fact that Martin had found Orbitoides 
in the marls of Cebu Island of the Philippine group, and declared them 
to be the Equivalent of the "Java Gruppe'^ which is Miocene, Smith 
referred the limestones of the Binangonan locality to the Miocene. 

The conspicuous outcrops of the limestone are, as has been stated by 
most writers, ridges or hills of a massive, whitish to yellowish, dense and 
fine-grained limestone showing evidence of metamorphism and usually 
breaking with a conchoidal fracture. 

The so-called Binangonan locality which is about 7.5 kilometers to 



80 ADAMS. 

the northeast of that town, is best referred to the barrio of San Guillermo, 
which is situated on the east side of the valley of the Morong River, 
opposite Riehthofen^s locality, which is on the west side of the valley. As 
seen from San Guillermo the limestone forms a high bluff with a peak 
rising to an elevation of about 100 meters above the wide, cultivated 
river valley. To the south of the peak the limestone at a much lower 
elevation disappears under a basaltic agglomerate. The matrix of this 
agglomerate i^ tufaceous and incloses fragments and even large blocks 
of cellular basalt. This agglomerate partially overlies the limestone on 
the west. The eastern slope of the limestone ridge is that of a valley 
slope with a talus of limestone. 

The limestone bluff extends northward to the vicinity of Teresa, but 
there it is not so massive. To the north of Teresa between two branches 
of the stream valley there is a rugged hill consisting of massive limestone. 
Following the trail from Teresa to Antipolo, limestone bowlders and 
outcrops are seen along the border of the valley and them in ascend- 
ing the hill argillaceous limestone strata and some beds which are 
even arenaceous are observed, but these exposures are- very limited, the 
country being covered with cogon grass. The limestone in the hill 
between Antipolo and Bosoboso is the most conspicuous outcrop in that 
vicinity, but it continues to the north and south in the ravines and on 
the hill slopes. In traveling to it from Antipolo one finds outcrops of 
a thinner, bedded limestone in descending from the upland on which 
Antipolo is situated^ so there is evidence of a zone in which this 
formation occurs. 

Returning to the consideration of the "fine-grained hornblendic 
slates'^ which Meyer saw lying on, the limestones and the "compact 
breccia composed of limestone and diabase-aphanite'^ which Yon Drasche 
observed dipping to the east on both sides of the limestone at Poray 
Creek and containing certain beds which he called diabase tuff, it is 
proposed to explain their characters by giving the succession at the 
locality of the gorge. Approaching the gorge from along the road from 
Montalbon, in the cuts in the foothills on the south side of the river 
there are exposures of basalt which has been sheared and jointed by 
d3mamic action and are considerably altered. The next exposures in the 
north bank of the river bed are variable, stratified, clastic beds which dip 
to the eastward. Succeeding these beds is a heavy limestone showing 
in a hill on the north side of the river and found at the roadside south 
of the stream. It is followed by stratified deposits which are sandstones, 
tufaceous beds, fine-grained, hard shales exhibiting color banding, fine 
and coarse conglomerates containing in places limestone pebbles, and . 
some brecciated beds. These exposures are in the bed and bank of the 
river and in the bed of a small tributary from the south. The dips are 
to the eastward and are steep. Succeeding the variable series just 



GEOLOGIC RECONNAISSANCE OF\SOUTH WESTERN LUZON. 81' 

described is the massive limestone in which the gorge is cJut. There 
are no good exposures just to the east of this limestone, the slopes being 
covered with a heavy talus ; but the first rocks in place were found -in 
the river bank are dacites. The contacts in the section just described 
are not plain. There has been severe dynamic action which has faulted 
and brecciated some of the beds and all are more or less altered. 

Coal has long been reported from the hills east of the town of San 
Mateo and to the south of the waterworks gorge. The beds are, how- 
ever, too thin to warrant exploitation. Kecently some prospecting has 
been done and the writer has been told that the coal occurs with some 
arenaceous and argillaceous beds which are associated with the limestone. 

Limestone is found near Angat. Itier, according to a citation by 
Yon Drasche, mentions limestone on the banks of the Angat Kiver, with 
, uplifted vertical strata and inclosing fossils. 

Tlie occurrence of limestones near the springs at Norzagaray to the 
south of Angat was noted briefly by Centeno, as was also an occurrence 
to the east of San Migwel near the springs of Sibul. The latter locality 
is just to the north of the area here under discussion. Both of these 
limestones are probably continuations of those near Angat. 

JMcCaskey has described more fully the outcrops near Angat. His 
localities are Mount Pecote, the banks of the Bayabas Kiver between 
Sampaloc and Bayabas, Bocol Hill to the south of the river and the 
Baras-Bacal Hills to the north. He states that the conspicuous lime- 
stones occur as massive beds overlying thin limestones, sandstones and 
shales. Mount Puning, in the Baras-Bacal Hills, is a conical hill of 
limestone through which the Santol Creek passes in a cave about 1.5 
kilometers long. In the shales associated with the limestones McCaskey 
found inferior thin beds of lignite, one at the barrio Sampaloc in the 
shales exposed by the Bayabas River, another in upper Sapa Santol, and 
a third in the Arroyo Laguio Malaqui near Norzagaray. The limestone 
areas were described by McCaskey but not mapped and are shown ap- 
proximately on the geologic map accompanying this report. 

Next in order may be mentioned the outcrops along the Mauban-Lucban 
Road on the southeastern part of the map. The occurrence of limestone 
in this section was suggested by the fact that Jagor and Roth described 
limestone conglomerates and conglomerates containing limestone pebbles. 
There is a massive limestone exposed in the river and near the road 
about 9 kilometers west of Mauban. Other outcrops probably occur in 
the section, since lime is burned from rock obtained to the east of Lucban. 
The presidente of Sampaloc reported the occurrence of limestone at 
several places which could not be located on the map because of deficient 
geographic details. It may be noted here that he described an in- 
termittent spring near Sampaloc which may prove interesting to anyone 
having time to study it. There are a number of exposures of shales and 



82 



ADAMS. 




' 



o 1 


!5" 




o 


«/)o J 


M 








m 


h 1 


u 


-^ O ^ 




«/) ro ^^ 






o 


Soil 





eS 






C 
< 






0) *J 

> '^ 

O 



<> 



AVNVl 






I- 






since the river 



sandstones in the cuts along the road between Mauban 
and Lncbanr to the east of Sampaloc^ where the road lies 
in the valley. The highland seems to be largely covered 
with tuff. These shales^ sandstones and limestone ap- 
parently belong to the Miocene series and have been 
folded and faulted so that the dips are variable. 

There is evidence of limestone outcrops near Pagbi- 
lao, since the stream there has brought down limestone. 
Just before reaching the Bagsabagsan Eiver on the road 
from Pagbilao to Atimonan, a massive outcrop of lime- 
stone is exposed near the road and forms a high bluff. 
Good exposures of sandstones and shales of varying char- 
acter and in part tufaceous are seen in ascending the 
zigzag to reach the summit section of the road. In de- 
scending to the east a massive limestone is found in a 
ridge striking about N". 30° W. The two limestone out- 
crops just mentioned are in every way comparable in 
importance and appearance with the typical massive lime- 
stone of this Cordillera. Other limestones quite similar 
but not so conspicuous were seen at points about 6 and 
2 kilometers west of Atimonan. There are additional 
outcrops of sandstones^ conglomerates and shales along 
the coast at Atimonan, and several occurrences of coal, 
probably not exploitable, are reported. The structure of 
the Cordillera in this section is that of close folding and 
faulting. 

Ickis crossed the eastern cordillera from Infanta (Bi- 
nangonan de Infanta) to Tanay on Bay Lake and con- 
tributed much to our knowledge of this region, previously 
geologically unexplored (fig. 5). Leaving Infanta he 
passed over an area of alluvial deposits. Following up 
the Agos Eiver he crossed a belt of andesites and then a 
belt of limestones, shales and graywackes in various stages 
of metamorphism, beginning a short distance above the 
Macadata Eiver and continuing along his route up the 
Alas-asin branch. He also reports the occurrence of coal 
of doubtful economic value found in this belt. The 
limestone contains, besides other fossils, Orbitoides as 
determined by Smith. The sedimentaries exhibit steep 
dips, folding and faulting and the strike is with the 
cordillera. To the west of this belt Ickis reports ande- 
sites and basalts probably extending into Jalajala Penin- 
sula and the basalts and tuff near Tanay. Some lime- 
stone probably exists to the west of the andesite belt, 
at Tanay brings dovm limestone pebbles. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 83 

The occurrences of limestones and -shales seen on an excursion from 
Bosoboso to the iron deposits near Santa Inez remains to be mentioned. 
On approaching Bosoboso some indistinct limestone outcrops were found 
in descending the hill into the town. In the river bed to the east of the 
abandoned settlement of San Jose and at the hot spring described by 
Abella as the "Mainit de Bosoboso" outcrops of somewhat metamorphosed 
argillaceous beds were seen which Abella has described as follows: 
"The country in these localities consists of an extensive formation of 
old slate; diabasic conglomerates (at Langay-langayan) and limestones 
(Lanatin Eiver). The strata at the hot spring strike N. 10° W. with 
dips very pronounced nearly vertical toward the east." He explains that 
the jointing and the fracturing of the formation is due to the igneous 
rocks which are found as dikes and volcanic masses. 

There is no true slaty structure in the formation at the hot spring or 
near by and it is evidently a mistake to class these rocks as slates or 
older slates, or to include them in the discussion of the older rocks as 
did Becker on the strength of Abella^s description. Certain hand speci- 
mens from the exposures at the Montalban Gorge are identical in char- 
acter with specimens from San Jose. The description by Ickis of the 
formation on the Alas-asin in his section from Infanta to Tanay shows 
that similar rocks are encountered there. 

Von Drasche in crossing from Laguimanoc to Atimonan found at first 
tuff, higher up tuff inter stratified with breccia of coral limestone and 
finally single isolated coral reefs which are rich in fossils. He says that, 
in part the limestone is completely crystalline. On the area of schist 
which he found he saw an isolated, small coralline liniestone cliff. In 
as much as Yon Drasche has seen the typical Tertiary limestone it would 
be fair to suppose in view of his description of the section that the coral 
reefs above-mentioned are not to be confounded with it, and that the 
breccia of coralline limestone with interstratified tuff beds is a formation 
distinct from the typical Tertiary, but after having seen the variable 
character of the Exposures at the Montalban Gorge and having studied 
the section from Pagbilao to Atimonan it seems more reasonable to accept 
the opposite conclusion. 

In this connection should be mentioned the conglomerate containing 
limestone pebbles described by Jagor and Eoth which the former found 
on the road from Mauban to Lucban. The locality which is especially 
described by Jagor is about 2 kilometers east of Sampaloc. When the 
writer passed this place a landslide had come down the hill to the river 
and carried away the road, leaving the formation well exposed. The 
limestofie pebbles which appear to have been derived from the dense 
Tertiary limestone of the region were seen lying in lenticular masses in 
a coarse-grained, soft sandstone-like matrix and dipping at steep angles. 
The matrix has been largely derived from the erosion of igneous rock 



84 ADAMS. 

and is in part tufaceous. The character of the conglomerate indicates 
that the older portion of the Tertiary formation had heen elevated and 
perhaps somewhat folded and subjected to erosion and that materials 
derived from it contributed to the formation of the conglomerate. Sim- 
ilar conglomerates were found at three places nearer to Mauban in cuts 
on the road. At one of these localities basaltic pebbles were mixed with 
limestone pebbles and at another the matrix was tufaceous. The dip 
of these exposures was not ascertainable. The district to the east of 
Lucban lies near the base of Banahao which is an extinct volcano and 
tuff deposits cover and obscure the structure of the adjacent upland 
portion of the eastern cordillera. 

At the locality of the iron deposits near Santa Inez on the Lanatin 
Eiver, bowlders of limestone are seen in the' river and on the hillsides. 
A massive limestone occurs on the mountain which contains the iron 
deposit and rests on diorite. About three hours walk up the Lanatin 
Eiver there is a mountain exhibiting a conspicuous white face of lime- 
stone. To the northwest of San Isidro there is a rough limestone hill 
which is conspicuous when traveling toward Santa Inez. 

It is safe to conclude that the limestones and associated sedimentaries 
of the eastern cordillera are Tertiary. Smith has recently referred 
them to the Miocene on the strength of having found Orhitoides in the 
limestones. Later, he has submitted a suite of fossils to a close com- 
parison with studied collections, and identification of the species points 
. to their being Oligocene. He will soon publish the results of this study. 

In passing it may be well to note that the fossil which McCaskey 
found near Angat and reported to be a staminate cone of a Lepidodendrid 
has been studied subsequently by the botanists of the Bureau of Science 
and they pronounce it to be of questionable value and certainly not a 
Lepidodendrid. 

The upland lying in the district defined by Luisiana, Caventi Majayjay 
and Pagsanjan and in which the rivers flow in deep gorges and where 
the Pagsanjan and Botocan falls are located, has a thick deposit of tuff 
and conglomerates which appear in places to be water laid. This 
deposit borders and in part overlies the folded sedimentaries of the 
eastern cordillera. It was impossible to determine 'definitely whether 
there is any water-laid tuff on the upland areas, which have an elevation 
of about 500 meters. The tuff undoubtedly occupies depressions. The 
youngest beds seem to be of sub-aerial origin. Between Mauban and 
the Mauban River there is a small hill in the recent alluvial deposits 
which is composed of water-laid tuff. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 85 

LOBOO MOUNTAINS. 

This region was observed from the vicinity of Batangas and from 
steamers which passed very near to its southern coast. In 1905 Smith 
made a partial reconnaissance of the Loboo Mountains, his special object 
being to examine some copper prospects on the Calbasahan River in the 
northeastern part of the area. He described the mountains as forming 
a chain approximately parallel with the coast but broken by the Loboo 
or Eosario River "Valley. From a study of the topography of the region, 
it seems that two ranges may be distinguished extending in a northeast 
southwest direction (IST. 60° E.) approximately parallel with the Rosario 
River Valley. If either of these interpretations of the trend of the 
Loboo Mountains is correct, the structure of the region is discordant 
with that of the eastern and western Cordilleras, and because of this it 
is discussed here under a separate heading. 

Diorites. — The only igneous rocks identified by Smith were diorites 
which he collected near the copper prospects in the northeastern part of 
the mountains. This locality would be considered as in the southeastern 
range if two ranges are distinguished. He found the rock to vary from 
a typical diorite to granitic and gneissic phases, being usually coarse 
grained. It is cut by quartz veins, more or less mineralized, and showing 
stains of iron or copper* oxide and carbonates.' Some molybdenite was 
found in the veins. Inasmuch as no important mineral deposits were 
developed by the prospecting, interest in the district ceased and there 
has been no occasion to revisit it. 

Andesites. — Beach gravels and bowlders from near the Pinamucan 
River were seen at Batangas. They had been brought in hancas and 
were being crushed for road metal to be used in the building of the 
road from Batangas to San Juan. These bowlders are contributed to 
the beach by the Pinamucan River and the other streams near by, which 
drain the western end of the northeastern range of the Loboo Mountains. 
They are andesitic and indicate the existence of extensive areas of andesite 
in the mountains. 

Tertiary sedimentaries. — In journeying from Batangas to the Loboo 
River Valle}^, Smith found exposures of what he described as a basal 
conglomerate succeeded by shaley sandstones and limestones. Fossils 
collected from this formation were determined by him as being of 
Miocene age and younger, certain of them having Bornean affinities. 
Accordingly, the sedimentary formations of the Loboo Mountains may 
be the equivalent of those in the eastern cordillera. 

Coralline limestones and marls, — Heavy beds of coralline limestone 



86 ADAMS. 

are exposed in the sea cliffs of the table-land at Locoloco Point and at 
Malabrigo Point. It is probable that this formation, which also contains 
some marls, is similar to the raised coral reefs in Santiago Peninsula, 
described subsequently in this report. The formation is perhaps the 
equivalent of the tuffs lying to the north of the Loboo Mountains and 
found to a limited extent within their area. The occurrence of the 
tuff formation in the Loboo Mountains has been noted by Smith and the 
writer has been assured by travelers that it borders the region to the 
north and that the occasional outliers are found within the mountainous 
district. 

Geologic history. — The occurrence of diorites in the Loboo Mountains 
and the existence of Tertiary sedimentaries and andesites indicate that 
the geologic history is complex and somewhat similar to that of the 
eastern cordillera. The relations of this mountainous area to the struc- 
ture in other parts of Luzon is discordant. Perhaps when the history 
of Mindoro and Verde Island, which intervenes, are well knowTi the 
Loboo Mountains may be found to be related to them. At present, the 
region is too little studied to warrant very definite statements regard- 
ing it. 

SOUTHWESTERN VOLCANIC BEGION. 

Extinct volcanoes. — There are three peaks in this region which are 
usually considered to be extinct volcanoes, viz, Banahao, Cristobal and 
Maquiling. Banahao and Cristobal probably should be referred to the 
same volcanic center, Cristobal being a subordinate cone. 

Banahao as seen from the southwest is quite symmetrical. It contains 
a large crater reported to be 210 meters deep. To the northeast there is • 
a sharp subordinate peak called Banahao de Lucban, or Banahillo. Cris- 
tobal lies to the northwest. It has a crater in which there are three 
small lakes. To the southwest there is Mount Masalacot and some hills 
which belong to the same area. There is no evidence of recent volcanic 
activity of Cristobal and eruptions attributed to Banahao should prob- 
ably be explained as avalanches of mud and stones produced by the break- 
ing of natural dams formed at the outlet of the crater, by landslides 
and debris falling from the cliffs. Jagor in a footnote states that it is 
recorded in the Estado Geogrdfico, Manila, 1865, that Banahao has been 
extinct since 1730, in which year it had its last eruption, breaking out 
on the southern part and vomiting torrents of water, hot lava, and 
stones of monstruous size, the signs of which may be seen until the 
present time at the town of Sariaya. This supposed eruption was also 
noted by Becker in his paper. In January, 1909, the newspapers stated 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 87 

that Banahao was supposed to be in eruption. Arriving at Sariaya it 
was found that an avalanche of water, mud and stones had descended to 
the west of the town along the drainage channels which are the natural 
outlet of the crater, and had covered the road with a layer of debris 
which in places was more than a meter thick. The water from the 
avalanche, heavily charged with sediments, reached to the sea, and in its 
descent changed some of the former channels and swept away a newly con- 
structed steel bridge on the Sariaya-Candelaria Eoad. As there was 
nothing which indicated a true eruption, no ascent of the mountain was 
made, but from the reports of others who followed the avalanche to its 
source it was learned that a landslide had dammed the chasm through 
which the crater has its outlet and that the bursting of this obstruction 
had liberated a lake which had formed therein. Another member of 
the Bureau of Science ascended the north slope of the mountain a short 
time previously. He stated that a lake had actually existed in the crater. 
It seems very probable that the catastrophe recorded as the eruption of 
1730 Avas, similar in every way to the one which occurred in this year. 
The velocity o»f the water which descended enabled it to carry great 
bowlders, some of which were fully 2 meters in diameter. -The mud in 
places covered cultivated ground and the force of the moving mass of 
water and stones uprooted trees and partially destroyed some coconut 
plantations. The town of Sariaya was located previously further to 
the west. The inhabitants moved to the present site after the avalanche 
of 1730 which destroyed most of their homes. 

The flanks of Banahao and Cristobal are covered with a heterogeneous 
mass of stones and detrital material which has gravitated down the slopes. 
Deep stream channels cut in this material are seen while driving along 
the road from Sariaya to the town of Tayabas, and likewise in passing 
around the north base of these mountains. Solid rocks are exposed in 
a veiy few places. At the foot of the mountain slopes this material 
gradually thins out and volcanic tuff forms the plain. 

Mount Maquiling has been made the subject of a special monograph 
by Abella. He states that its summit has somewhat the appearance of 
a crater broken dow^n by erosion. The positions of the peaks which form 
its summit when seen from the southwest near Santo Tomas, as well as 
the outline, suggest what was probably once a distinct crater. There are 
no references to an eruption of Maquiling, but the mountain contains on 
its slopes and around its base many signs of expiring activity. The hot 
springs near Los Bafios and the so-called solfataras or hot springs on 
its flanks may be especially mentioned. The positions of the hot springs 
and solfataras as recorded by Abella will be seen in the accompanying 



88 



ADAMS. 



illustration {^g. 6) which is taken from a map accompanying his report, 
and which describes them in detail. Most of the solfataras are in the 
nature of hot mud spots, from which emanate more or less boiling water 
and steam. Their action has converted the surrounding rocks into a 
white earth similar to kaolin and deposits of this earth are also found 
on the slopes of the mountains at places where solfataric action has 
ceased. The most important evidence of volcanic activity near the base 
of Maquiling is the crater lake called Laguna de los Caimanes which 
probably occupies the crater of an e:xtinei cinder cone. It lies on the 
border of Lake Bay just to the west of Los Bafios and by some author- 



B A Y LAKE 

C? Calambal. 




LEGEND 
EUlDoleritesCIZlTuff and lapilli 
S Solfbtaras T Thermal springs 



piQ, 6, — Sketch Map of the Geology and Topography of Mount Maquiling 
AND the Surrounding Country, as Mapped by Abella. 

ities is described as an island. It appears that during the dry season 
a crossing on dry land to this crater-lake is feasible by avoiding the 
hot springs which intervene between it and the mainland. During the 
yainy season the rise in level of Lake Bay converts the cinder cone into 
an island. A small hill, Pansol,, to the west, at the base of which 
issue hot springs, and a hill to the east on Point Mayondon, are probably 
remnants of other cinder cones. A small, low hill called Cerro de la 
Mesa is seen when traveling from Calamba to Los Banos by wagon 
road, to the northwest of Maquiling. This appeared upon examina- 
tion to be a gmall cinder cone preserving some signs of a crater. Sur- 



* GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 89 

rounding it there is a crescentic ridge which probably represents a part 
of the older outer crater {^g, 7). The region around the base of Ma- 
quiling has not been thoroughly studied, but some of the hills to the 
west appear to have been formed by subsidiary vents and one to the 
west of the road between Calamba and Santo Tom4s, which was ex- 
amined, gave evidence of being a cinder cone, but without any indica- 
tions of a crater. 

Eruptives. — Under this head will be discussed the mountains of the 
southeastern volcanic district which do not retain evidences of distinct 
craters or are not so close to the extinct volcanoes as to appear genet- 
ically related to them. 

To the south, on the structural line which includes Talim Island and 
Maquiling, stands the elongate Malarayat-Sosoncambing Mountain. To 
the east of Maquiling and to the northwest of Banahao there is a group 
of low mountains to which here is given the name Calauang Mountains. 
Further east in the tuffs near Cavinti and to the northward .there are 
some outcrops of igneous rocks which are not shown on the geologic 
map. A small hill occurs at Pagsanjan and to the east of Pagsanjan 




Fig. 7. — Sketch Showing the Extinct Cinder Cone Cerro de la Mesa and 
Outer Crater Rim as Seen Looking South from Calamba Railway Station. 

the country contains some topographic features which indicate eruptives. 
Southeast of the Pico de Loro Mountains there is a group of peaks 
which are surrounded by the tuff formation. To the south of Nasugbu 
some exposures of igneous rocks were found which belong to the sub- 
structure of the Santiago Peninsula. In the peninsula south of Lemery 
there is another group of eruptives. 

Finally, there should be mentioned Mount Sungay to the north of 
Taal Lake, Mount Butulao to the west of the latter, but a considerable 
distance from it, and Mount Macalod standing on the southeast margin 
of the lake and having a high precipice facing it. The relation of these 
eruptive masses, with the exception of Malarayat-Sosoncambing which 
is considered to be on the north-south structural line that includes 
Maquiling and Talim, is far from clear. It appears that the eruptives 
lying in the peninsular area have been added to the mainland of Luzon 
Island by the gradual extension of the tuff formations. The Calauang 
group of lower peaks is very irregular. The igneous rocks further to 
the east near Cavinti and Pagsanjan are nearly covered by the tuff 

93217 3 



90 ADAMS. ^ 

formation. The group of mouDtains southeast of the Pico de Loro 
group has not been very thoroughly studied and it is possible that it 
should be considered as related to the Pico de Loro group^ but in this 
report it is considered with the southwestern volcanic region, because 
of the finding of extensive masses of basalt near the southern bases of 
the mountains of the group, which class of rock is rare in the western 
Cordillera. 

The three peaks which are found near the border of Taal Lake have 
been thought by some to be subordinate cones near the base of a supposed 
former lofty cone of Taal, but the writer does not incline to this idea. 
If they are related to a common center of eruption, they form a much 
larger group than do the peaks related to Banahao. Careful petrographic 
study might show whether or not these mountains have had their origin 
from the same magma. Macalod has the apperance of an independent 
eruptive mass, with subordinate peaks, one of which is clearly defined 
to the southwest of the main mountain. 

Cinder cones. — From the descriptions of the historic eruptions of 
Taal Volcano and an examination of its crater it is evident that in 
recent time no lava was ejected in the form of flows, but that bombs, 
scoria and ashes have been thrown out. In other words, Taal Volcano 
is a large cinder cone. A number of craters which are extinct occur on 
Taal Island and examination of them proves that when they were in 
eruption they likewise were of the cinder-cone type. 

Within the area of the tuff formation which has been built up from 
the volcanic ejectamenta of the southeastern district, there are some 
isolated cinder cones which still preserve indications of craters and no 
doubt several of the smaller hills are the remnants of cinder cones which 
have been largely eroded. South of Lipa and near the road to* Kosario, 
there is a hill which has a flight depression in its summit, and around 
this depression there is a crescentic rim on which scoriae are found. The 
slope of the hill is covered with scoriae and lapilli. 

Just to the north of New Eosario there is a crescentic-shaped hill 
which is a remnant of a cone. The hill is broken down to the northeast, 
Masses of scoria outcrop on the higher part of the crater rim. A hill 
which retains some of the characteristics of a cinder cone is situated to 
the northwest of Batangas, but the rim and the crater are not very 
well defined. To the west of the base of Maquiling there are a n amber 
of hills which may represent small crescentic cones. About half-way 
between Calamba and Santo Tomas and to the west of the road there 
is a subconical hill formed of lapilli. This may be an eioded cinder 
cone. The hill containing Laguna de los Caimanes, Pansol, the hill 
on Mayondon Point ' and Cerro de la Mesa have been mentioned as 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 91 

probably being cinder cones in describing the evidences of activity around 
the base of Maquiling. 

Crater lakes. — The crater lakes probably owe their origin to phenomena 
closely allied to those which produced the cinder cones. The rims of 
the lakes usually rise a little above the surrounding country and suggest 
that they are the remnants of cones which have collapsed and subsided. 
In the country around San Pablo there are a number of crater lakes 
which are readily accessible, and San Pablo is coming to be known as 
the center for tourists visiting the crater-lake country. The first descrip- 
tion of these lakes is by Jagor, who visited several situated near San 
Pablo. He described them as having circular embankments with very 
gentle outer slopes formed of lapilli. The interior walls of the rim are 
usually very steep. Most of the lakes are drained through clefts in the 
crater border which have been formed by erosion. The streams flowing 
from them, like most of the streams of the volcanic region, have steep 
banks and narrow channels. The lake most easily accessible is the one 
situated just north of the town of San Pablo and it may be reached by 
a few minutes' walk to the end of one of the streets. Two lakes occur 
on the public highway east of San. Pablo. They are near the road and 
can easily be visited by travelers by stopping a few minutes and walking 
through the groves of coconut palms. 

These lakes, besides being of interest because of their origin, are made 
picturesque by the vegetation which grows on their borders. Some of 
them afford a good supply of fish to the neighboring inhabitants. 
However, others are partially drained and the abandoned lake beds are 
planted with coconut trees. Not all of the crater lakes which have been 
indicated on the reconnaissance map were visited, but information 
concerning them has been obtained through inquiry. It is interesting 
to note that a lake which occupies the remnants of a small crater is 
situated near l^asugbu. The distribution of the crater lakes and cinder 
cones shows volcanic activity to have been extended over a large part 
of the southeast district and it is probable that the deposition of the 
tuff formation has obscured many of the early vents and that others 
have been completely eroded. 

The active Volcano Tool. — Taal Volcano is situated on an island in 
Taal or Bombon Lake. The island, on which are found a number of 
extinct cinder cones and the active crater, has been built near the center 
of the lake by late volcanic activity (Plate III). The main crater which 
is situated near the center is usually referred to as Taal Volcano. It 
is approximately circular in form. The southwestern border of the 
crater rim rises to an elevation of 320 meters, which is the highest point 
on the island. The lowest points on the rim are about 130 to 150 meters 



92 ADAMS. 

in elevation. The lowest points on the floor of the crater are about on a 
level with the water of Taal Lake. Visitors usually ascend the eastern 
side of the crater rim where it is low and near the shore; the most 
desirable time to make the ascent being about dawn and the rim should 
be reached by sunrise. The descent into the crater is by a steep zigzag 
path, but presents no serious difficulties. During the first cool hour or so 
of the morning the vapors lie low over the lakes in the crater. On 
sunny days as the sun warms the air, the steam begins to ascend and 
the vapor and sulphur fumes are wafted about so that the view is not 
so clear and travel in some parts of the crater is made difficult and even 
dangerous. When the atmosphere is warmed by the sun, the steam from 
the green lake usually rises and forms a mushroom shaped cloud which 
soon becomes detached from the column of vapor which forms, its stem. 
As the heat of the day advances the vapors become more attenuated and 
are blown about by the rising wind, but before this time the visitor 
should have left the crater if he wishes to avoid an arduous climb in the 
hot sun. On days when the sky is cloudy, the crater may be explored 
more at leisure. 

Two lakes lie within this crater. They are usually called the yellow 
lake and the green lake. During the rainy season there is a third tem- 
porary red lake. The yellow lake receives the natural drainage of the 
crater. It appears to be shallow and is hot, but does not boil. The green 
lake gives off steam from its surface and near its southern border boils 
violently as if over a vent. A circular crater is located to the south of 
the green lake. On its floor there are several boiling mud spots from 
which but little vapor rises. On the south border of the yellow lake 
there is a cone, called the red cone, because of the color of its crater. 
It is broken down on the south side and drains around its eastern base 
into the yellow lake. A vent from which steam issues with great force 
occurs on its northern outer base. The yellow lake now extends to this 
vent but formerly was separated from it by a narrow isthmus. There is 
a remnant of an older, large crater rim which forms a crescentic ridge 
rising southeast of the yellow lake and curving around to the south of 
the green lake, passing between the green lake and the crater with the 
mud spots. Visitors to the volcano usually travel on this ridge when 
exploring the crater of Taal, since it affords a fine view and an easy 
means of approaching the principal points of interest. The ridge is 
very narrow between the green lake and the crater containing the mud 
spots and care must be taken that the wind is not blowing the steam and 
sulphur fumes in such a direction as to cause them to inconvenience or 
overcome the visitor. There are small cracks in this ridge from which 
sulphur fumes issue and at its western end, which is south of the green 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 93 

lake, a descent and detour may be made to the south to reach the 
remnants of some extinct cones which have level floors. Sulphur fumes 
and a little steam rise from the cracks in their walls and floors. 

The layers of tuff exposed in the inner wall of the outer crater exliibit 
bedding and show a banding of colors varying from whites and grays 
to yellows and reds. The colors are especially vivid on the walls near 
the points of activity where the hot acid vapors are sublimed on the 
rocks, forming an efflorescence of iron, aluminium and magnesium sul- 
phates and chlorides producing variegated white, yellow, red, blue and 
green effects. Small amounts of native sulphur occur in crevices and 
cavities, but it is inconspicuous. 

In 1880 Centeno visited Taal Volcano while studying the effects of the severe 
earthquake of that year, and found that the volcano showed no signs of in- 
creased activity. From his sketches of the appearance of the volcano and his 
description we learn that there was at that time one active cone sending up a 
great quantity of vapor. This active^ cone seems to correspond with that ^of 
his description written in 1882 and is now the crater containing the mud spots. 
Near the active cone there were some smaller ones which are shown on one of 
his sketches as giving off thin columns of vapor. 

The yellow lake is described as containing a place near its center where the 
water boiled up violently and occasionally threw up black mud and sent out 
waves which agitated the surface of the lake. 

In 1882 Centeno made a study of Taal Volcano and his report is accom- 
panied by a small hachured map which shows most of the features quite 
clearly. A model of the volcano inherited from the Inspeccidn de Minas, is in 
the division of mines. It is probable that the model was based on the map or 
was made at about the same time. Studying the model, which is somewhat 
clearer than the map, but one defect is found which can not be reconciled with ♦ 
the present features of Taal or be accounted for by subsequent changes. The 
model shows the crater of the red cone as being broken down on the side toward 
the green lake. As it now exists, it is broken down to the south and drains 
around its eastern base to the yellow lake. 

Making due allowance for the imperfections of the model and map. 
certain changes have taken place since Centeno described the volcano. 
The green lake has extended its area to the southward and contains a 
point where there is violent boiling. The neck of the peninsula which 
extended into it from the east has been eroded, leaving the point as a 
pyramidal rock or island in the lake. A steam vent and a sulphur 
dioxide gas vent separated by a distance of about 15^ meters and evid- 
ently not connected near the surface have formed at the north base of 
the red cone. The small inactive cones G and H of Centeno^s descrip- 
tion have been largely obliterated by erosion. The active cone which 
was described as being about 15 meters high, quite perfect and the point 
of most activity, giving off great masses of water and vapor, is now 



94 ADAMS. 

broken down, but there are some vents in the remnants of its walls. 
The inactive small cone F of his description has disappeared. 

In 1904 Taal was reported to be in eruption and many people visited 
it then and photographed it. Dean C. Worcester, Secretary of the 
Interior, has some good negatives taken during this active period and 
they furnish a better idea than descriptions of what took place during 
July, which, . stated briefly, was as follows. 

Besides giving off a greater amount of steam and sulphurous vapors 
at all points where such actions had usually occurred, the new crater 
formed in 1904 sent up great columns of vapor and intermittently ejected 
mud and stones to a height of at least 150 meters. The activity of 
this crater soon diminished and in December, 1905, its floor was covered 
by a boiling lake. In March, 1907, the lake had disappeared. The 
crater now contains some hot mud spots which are in a state of very 
moderate agitation and give off but little steam and gas. The temporary 
red lake in the southeastern part of the main crater floor seems to have 
first been mentioned during the eruption of 1904.^ 

Historic eruptions of ToaL— The following summary of the historic erup- 
tions of Taal is taken from Centeno's monograph published in 1882. At the 
time of the Spanish conquest, according to various ancient documents, the place 
of activity of Taal (or perhaps better, a place of activity) was on the north- 
western point of the island in a small cone, Bininting Malaqui. There are no 
records of eruption from it, but in 1880, when it was visited by Centeno there 
were some small vents of vapor and gas in its crater and one on the southeast 
border of its rim. There are now some similar phenomena at this cone, but 
of diminished importance. 

In 1680 Taal is described as having occasional eruptions which destroyed 
many fields on the island. At that time there were two vents, one of sulphur 
and the other of green water. They are interpreted sls being the yellow and 
green lakes which exist to-day. 

In 1709-1715 eruptions produced some damage on the island. 

In 1716 there was an eruption accompanied by earthquakes, and a disturb- 
ance in Taal Lake between the island and Mount Macalod. The waters of 
the lake were so agitated that waves cut away the outer shore of the lake, 
injuring the convento at Taal which was then situated on its border. Tliere 
are some sunken rocks in the lake between the volcano and Mount Macalod, 
but whether or not they originated at the time of this eruption is not known. 

In 1731 there was an eruption in Taal Lake east of the island, forming a 

»In the article by R. F. Bacon, This Journal, Sec. A, (1907) 2, 115, on the 
"Crater Lakes of Taal Volcano," the yellow lake of Centeno's description is called 
the boiling crater lake. A temporary lake southeast of the yellow lake is referred 
to as the yellow lake. Evidently the colors of the lakes vary from time to time. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 95 

new island the larger part of jvhich subsided, leaving remnants which are known 
as Bubuing and Napayong Islands. 

In 1749 a violent eruption began on August 6 and continued about three 
weeks. Some settlements near the shores of the lake suflfered damage from the 
seismic disturbances and falling ashes. There were no flows of lava, but in- 
candescent stones were thrown out. Eruptions took place in the lake to the 
north and east of the island. 

In 1754 the greatest historical eruption took place, lasting from May until 
I>ecember. The towns of Sala, Lipa, Tanauan and Taal, then on the border 
of the lake, were destroyed, and Balayan, Bauan, Batangas, Rosario, Santo 
Tomas and San Pablo suffered great disasters. The rain of scoria and ashes 
extended over a much larger area. The surface of the lake was covered with 
patches of floating pumice and ashes. Incandescent stones were thrown out 
and fell into the lake. There were, however, no lava flows. Heavy rains 
descended and the barrio of Balili near Sala was covered with a shower of 
mud. 

In 1808 there were eruptions of slight intensity which began in February 
and continued until the end of April, but no damage was done to the inhab- 
itants who, after the disaster of 1749, had returned to settle on the margin 
01 the lake. 

The eruption of 1904 which has already been described was unimportant. 
It is the first one of which photographs have been published. 

Agglomerates of the southwestern region. — There are extensive beds 
of agglomerates found in the tuff formation, as has already been noted. 
They are seen principally around the northern base of Banahao where 
the streams flow in deep gorges. In other localities angular fragments 
of andesite and basalt are found included in the tuff but not forming 
true agglomerate beds. 

To the west of Balayan there is a line of sharp peaks which present 
an escarpment on the eastern side. The exposures which were seen in 
sheer faces in this escarpment appear like conglomerates. Masses which 
had fallen from the cliffs were examined and although the material was 
badly weathered it seemed that the formation should be classed as an 
agglomerate. Its occurrence is noted in the accompanying geologic map. 
It is probable that a fault passes along the- eastern base of this line of 
peaks. The origin and extent of the agglomerates and the position of 
the fault iine could not be determined. 

Tuff formation. — The area of the tuff formation, as shown in the 
accompanying geologic map, includes localities which have been described 
by previous writers as containing principally water-laid tuffs. The 
higher areas probably include subaerial formations and near Taal Volcano 
there are remnants of recent showers of volcanic ash. Von Drasche 
indicates in his map (Plate I) a large portion of the tuff deposits. 



96 



ADAMS. 



Centeno has also published a partial mapping, of the tuff deposit (^g. 8) 
which for the most part corresponds with the largest area shown on 
Yon Drasche^s map, but continuing further to the north. In the map 
accompanying this report andesite tuff and diorite tuff as distinguished 




Tuff 



^-^^<^^ 



Fig. 8. — Sketch of the Tufaceous Area of Taal Volcano, as Mapped 

BY Centeno. 

by Von Drasche are included in one general area. The area of trachyte 
tuff mapped by him in the central valley is here shown as alluvium. 

Taal Volcano has been thought by some to be the origin of the greater 
part of the tuff, but, as Becker has pointed out, there is reason to believe 
that it may have been derived from many sources and been gradually 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 97 

deposited during a long period of time. The formation varies from fine- 
grained, water-laid strata to coarse, breceiated deposits, agglomerates and 
conglomeratic phases. The examination of samples obtained from drill- 
ing deep wells shows interstratified marine sands and some clayey beds. 

It is very difficult to distinguish between the water-laid tuff and the 
subaerial deposits because of the similarity in the materials and the 
absence of exposures at critical points. On the divide between Bay Lake 
and Manila Bay in a railway cut some shark's teeth were found in a 
stratum containing waterworn pebbles. The bed is evidently a marine 
or estuarine deposit. The elevation of this point is about 40 meters 
above the sea. The presence of the fish teeth indicates an emergence 
which may be correlated with the change of level indicated by the marine 
conglomerates on the flanks of the western cordillera. Smith collected 
some fragments of a mammalian tooth near the Loboo Mountains at a 
considerable elevation, which, however, was not determined by measure- 
ment. He believed the tooth to belong to the Bovidae. An examination 
of the pieces which have been preserved showed that they correspond 
fairly well with the teeth of the Cervidae and do not differ much fi*om 
those of the living Philippine deer. This tooth may have been buried 
in late subaerial deposited tuff and the presence of plant impressions, 
especially of blades. of grass or rice in the same beds, is corroboratory 
evidence. The mammalian tooth found in the Pasig well at a depth of 
between 80 meters and 85 meters^ appears to belong to an extinct species. 
The formation may, however, be marine and it may have been buried in 
deep water deposits. The locality where the fish teeth were found and 
the situation of the Pasig well are not very widely separated. Outcrops 
of the tuff at intermediate localities appear to be water laid and contain 
many plant remains. 

The higher portion of the tuff area is a ridge between Mount Sungay 
and Mount Balulao called cordillera of Tagay-tay on d^Almonle^s map. 
The southern face of this ridge presents rather steep slopes and some 
escarpment, while the northern side of the ridge slopes gradually toward 
Manila Bay. It has been argued by some writers that this ridge re- 
presents a portion of the base of the former lofty cone of Taal Volcano. 
The elevation of the ridge varies from 500 to 600 meters, which is con- 
siderably in excess of the highest marine deposits on the flanks of the 
western cordillera. There are some basalts in the southern face of the 
ridge, but most of the exposures are beds of stratified tuffs. Certain of 
the beds appear to be water-laid. If any part of the higher beds in this 
ridge were deposited in the sea, it indicates a greater elevation of the 
locality than has taken place in the adjacent areas. It is not improbable 
that the ridge, together with two mountains . which form its determina- 
tions, is the southern portion of a fault block which has been lifted. 



98 ADAMS. 

The escarpment which runs northward from Mount Gonzales along the 
western limit of the alluvium which borders Bay Lake, and continues 
to the west of the Mariquina Valley may owe its origin to a fault scarp 
which has receded westward by erosion. When one studies the gap 
between Mount Sungay and Mount Maquiling, he is impressed with the 
fact that the country is much lower than the area to the west, and finds 
it difficult to acount for the discontinuance of the ridge eastward from 
Gonzales except by means of faulting. 

The general level of the country east of Gonzales Mountain and Taal 
Lake is continued westward between the Tagay-tay ridge and Taal Lake 
for a short distance and when viewed from Taal Lake has the appearance 
of a terrace. (See fig. 9.) This seems to indicate that the deposition 
of the tuff formation was continued after the elevation of the Tagay-tay 
fault block. 

On the eastern border of the tuff area, which is indicated in a very 
rough way on the geologic map, the deposits overlie the folded tertiary 
formations. In this locality it is likewise difficult to decide whether the 
tuff deposits are subaerial or water-laid and much further work must be 
done in order to determine their full extent. The tuff is found at 

_ ^ Mt. Gonzale& 



Fig. 9. — Sketch Showing the Relations op Tagay-tay Ridge and the Upland 
ON THE Northeast ' Border of Taal Lake as Seen from Lipa Point. 

levels of at least 50 meters above the sea and there is naturally some hesita- 
tion in assuming that there hais been so great an emergence as would have 
been required to lift the tuff beds from below sea level. 

In the upland to the southeast of Pagsanjan the streams flow in deep 
gorges in the walls of which coarse agglomerates with a matrix of tuf ace- 
ous material are exposed. In places there are clear indications of bed- 
ding and some strata of the tufaceous material are clearly water-laid. 
The escarpment at the border of the upland east and southeast of Bay 
Lake appears to have once been a cliff and later its base was the ])order 
to Bay Lake which has since receded from it. With the elevation ,of 
the area the streams of what it now the upland, cut their valleys deeper, 
forming gorges and the falls caused by the resistance of the harder beds 
of the agglomerates have gradually receded to their present positions. 

In this connection it is interesting to note that some tufaceous deposits 
related to the area of basaltic eruptives in the vicinity of Antipolo appear 
to be water-laid and contain conglomerate phases well exhibited in the 
stream bed near the railroad bridge. 

At the gorge of the Mariquina Eiver, where a dam has been built as 
the headworks of the Manila water supply, it is apparent that the river 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 99 

has cut its channel down through the ridge of limestone which forms 
the gorge. In the accompanying profile (fig. 10) traced from a photo- 
graph, the probable former levels of the stream are indicated by benches 
on the walls of the gorge. The elevations of these benches were de- 
termined with considerable accuracy by the engineer in charge of the 
construction of the dam. The highest is at a level of 342 meters. If 
these benches have been produced by the side cutting of the Mariquina 
Eiver it indicates a very considerable elevation of the locality and one 
which is commensurate with the elevation which may have taken place 
at Antipolo and in the southern part of the eastern cordillera. The 
Mariquina River, after passing the gorge, reaches the tuff formation 
near Montalban. The tuff being water-laid indicates the former presence 
of the sea at Montalban and at that time the river may have flowed into 
the sea but a short distance from the gorge. Subsequently, as the land 
was elevated, the river has cut its channel to the present grade. 

Elevation 390 m,---- 



D am foundation elevation 39 m^ 
FIG. 10. — Profile of the Gorge of the Mariquina River at the Head Works 
OF THE Manila Water Supply, Showing River Terraces Cut in Limestone. 

Coralline limestones and marZs.— Dana records in discussing the 
geology of Luzon that he was told of a bed of coral occurring at Point 
San Diego, at ^ elevation of 600 feet above sea level. Later, Von 
Drasche in writing on the Philippines, cited Dana aB authority for the 
occurrence of raised reefs, but seems to have made a mistake in the 
locality, since he says that they are found on Santiago Point, and on 
his geologic map he colored a portion on Santiago Point to show the 
presence of younger coralline limestones. Not having access to Dana's 
paper, and accepting Yon Drasche's mapping the writer made a trip 
westward from Balayan into Santiago Peninsula to- see the extent of the 
formation. Coralline limestones and marls were found at an elevation 
of 200 meters. This formation together with the surface soil and 
vegetation, effectively concealed the underlying rocks. The limestones 
have the appearance of reefs and occur at elevations intermediate between 
the highest on the upland and the living ones in the sea. Mr. Clark, 
who was my assistant in the field work, made an excursion south from 
Nasugbu and passed near San Diego Point. He reported the occurrence 



100 ADAMS. 

of marine sediments similar to the marine conglomerates found on the 
flanks of the Pico de Loro Mountains, but containing few bowlders and 
some coralline reefs. On the strength of these observations, Santiago 
Peninsula is mapped as occupied principally by younger coralline lime- 
stones and marls. 

. An area of coralline limestones also occurs southeast of the Loboo 
Mountains and, as has been explained in another part of this paper, 
these limestones are associated with some marls and probably grade into 
tuffs. 

In making an excursion south of Lemery along the sea beach to 
Point Ligpo, a limestone was found included by basalt and evidently 
metamorphosed by contact with the igneous rocks into a very dense 
stone. A peculiarity of this limestone not noted elsewhere in studying 
the similar formations, was the presence of small cavities, or druses, 
lined with clear quartz crystals. The exposure occurs on the beach and 
could not be followed very far inland. The peninsula was not thoroughly 
explored. It may be that this limestone is a fragment of a tertiary 
formation, but it seems more probable that it is a younger coral reef. 

Classes of igneous rocks. — Some petrographic work has been done upon 
the rocks of this district, but it has been confined to a few specimens 
collected during reconnaissance work. Oebbeke classed the rocks of Taal 
as volcanic augite-andesite. Yon Drasche recorded them a-s dolerite 
with certain exceptions which might be classed as augite-andesite. 
Becker says that the active and exinct volcanoes of southern Luzon 
appear to be mainly andesitic^ but not devoid of basalts. 

A number of specimens were collected for petrographic study during 
the field work which was carried on for this report. They have been 
determined as andesites and basalts, but there is a gradation from one 
class -of rocks to the other. Apparently the basalts are the younger 
rocks and they are found associated with the andesites in practically all 
the eruptive centers of the district. 

Structure of the region. — There has been a tendency on the part of 
former writers to consider the active and extinct volcanoes of this district 
as forming a part of a chain extending into southeastern Luzon. This 
is at variance with the structure of the island which, with the exception 
of the Loboo Mountains, has a northwest trend. 

Banahao and its subordinate peaks lie near the belt of folded tertiary 
sediments which have a strike of N. 30° W. and in so far as is 
known there is no great fracture crossing the Tertiary formations. It 
has been pointed out that Binangonan Peninsula, Talim Island, Mount 
Maquiling, and Mount Malarayat form a chain of eruptives on an 
approximately north and south line, or on two parallel lines having north- 
west trends. Taal Volcano and the eruptive peaks which are grouped 
around it are in strike with the axis of the central valley of Luzon. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 101 

Arayat, which rises as a lone peak in the central valley, probably lies 
on the same stru'ctnral line as Taah The volcanic centers of the south- 
west district are thus readily referred to a parallel system of fissures 
which corresponds to the trend of the east and west cordilleras. The 
location of the volcanic peaks of southeastern Luzon bear a similar rela- 
tion to the trend of the mountain ranges of eruptive rocks and the belts 
of folded sediments in so far as they are now known. 

The emergence of the marine conglomerates, water-laid tuff and 
coralline limestone has already been touched upon" in previous paragraphs. 
The amount of emergence is shown to be at least 200 meters in the 
weistfern cordillera district and there is some evidence that it is greater 
in the area of the eastern cordillera district, possibly amounting to as 
much as 500 meters. It is hardly probable that such great changes of 
level should have taken place without being accompanied by displace- 
ments. Two probable fault lines have been mentioned, one passing 
to the east of the agglomerate area near Point Santiago and the other 
extending northward from the base of Mount Gonzales and passing to 
the west of Bay Lake. The volcanic rocks and the volcanoes lie along 
great fissures and it is probable that displacements have occurred on 
these structural lines during the emergence of the region. 

It has been argued in this paper that the probable trend of the struc- 
ture is northwest with the exception that in the Loboo Mountains it 
is to the northeast. It has occurred to me that a structural line having 
a northeast trend may pass through Balayan Bay, Taal Volcano and 
the southeastern border of Bay Lake. If such is the case the line would 
form the southern border of the elevated fault block which includes 
Batulao, the Tagay-tay ridge and Gonzales. 

It is well known that southwestern Luzon has suifered severe earth- 
quakes. There has been a tendency to refer these seismic disturbances 
to volcanic centers. It would seem more desirable to refer them to 
tectonic lines ; but the systematic study of the earthquakes has not been 
carried on for a sufficient length of time to make this possible. 

Origin of Taal Lake, — Father Zuniga regarded Taal Lake as originat- 
ing in the collapse of a volcanic cone, and this theory was accepted by 
Hochstetter, Yon Drasche and Centeno. Becker says that the theory 
of volcanic collapse "seems to imply that an empty space beneath the 
earth^s surface is formed by the eruption of lava and that the inter- 
vening rock is too weak to bear the load put upon it. . . . I doubt 
this theory as applied to volcanic cones excepting only when invoked 
to account for local details of structure. ... On the other hand, it 
is well known that craters of vast size have been formed by explosions, 
and I can see no reason to doubt that Bombon may have been, probably 
has been, formed in this way in spite of its large dimensions." Taal 
Lake is evidently a caldera formed by peripheral and radial faulting 



102 ADAMS. 

and the subsidence and collapse of the many cones which have been 
formed within its area during prehistoric eruptions. The process of 
formation has been continued to a small degree during historic time. 

The fact that the caldera is occupied by a lake prevents in a large 
measure the study of its origin. However, in the brief notes concerning 
the two eruptions of Taal, it is recorded that within Taal Lake a new 
cone arose as an island and subsided, leaving as its remnants two small 
islands which lie to the east of Taal Volcano. In another eruption, a 
portion of the shore of Taal Lake near the former site of the settlement 
of Taal subsided below the water. Many who see the sheer face on the 
west side of Mount Macalod form the opinion that a large mass of the 
mountain has subsided into Taal Lake, and as substantiating this idea 
it is pointed out that the deepest part of the lake is found near this 
mountain. While studying the shores of the lake it was observed that 
in certain parts they are precipitous, while in others they are eroded 
into gentle slopes. It may be that some of the precipitous shore lines 
are due to recent faulting and displacement, but it will require a detailed 
study to prove this. Anderson,* in his article on Aso Volcano in Japan, 
has given a list of 31 important calderas, among which he mentions 
Taal. He points out three ways in which calderas may have originated, 
namely, by being built up around a vent of great magnitude, by the 
explosive removal of a volcanic mountain or by the subsidence of the 
area inclosed within its walls. The writer does not feel competent to 
discuss the theory that Taal Lake has been formed by explosions as has 
been suggested by Becker, but not fully outlined by him in its applica- 
tion to the volcanic center of Taal. 

There are many objections to Centeno^s elaboration of Zuniga's theory 
that Taal once had a high volcanic cone. Centeno assumed that the 
ridge (cordillera of Tagay-tay, elevation 500 to 600 meters) between 
Mount Gonzales and Batulao represents a portion of the base of a 
former cone which he reconstructed, theoretically assigning to it an 
elevation of 3,750 meters. The ridge, however, does not have the form 
of a portion of the base of a cone and moreover there does not remain 
any evidence of similar features to the east and south of Taal Lake, but 
instead, the tuff formation is nearly horizontally bedded' and forms a 
plain with an elevation ranging from 150 to 250 meters near the lake. 
A large part of it is certainly water-laid. It is manifestly incorrect to 
assume that Mount Macalod is a part of the southern base of a former 
cone of Taal as Centeno has argued, since it is an eruptive mass unlike 
the Tagay-tay ridge. It is suggested that the ridge to the north of 
Taal Lake may be due to elevation of a fault block. While there is 
little certainty that this is true, it would be consistent with the general 

* The Japanese Volcano Aso and Its Large Caldera. Journ. Qeol. (1908), 
16, No. 6/ Univ. Chicago Press. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 103 

structure of the region and would account for the topographic features 
north of Taal including the escarpment west of Bay Lake. As has been 
quite fully accepted by all writers, the sea once flowed through a strait 
between the eastern and western Cordilleras and at that time it oc- 
cupied a large part of the southwest district. There is no need to as- 
sume that Taal Volcano has ever formed a peak of great altitude in 
order to account for all phenomena connected with it, and the topography 
and structure of the region can be adequately explained if it is considered 
as a volcano which has always had it§ crater at a 'relatively low elevation. 

The outlet of Taal Lake is the Pansipit Eiver. This stream has 
shallow places in which large bowlders of basaltic rock are exposed and 
some of these seem to be in place. The Pansipit Valley appears to have 
been eroded by the stream which occupies it. At the outlet of the lake 
it is bordered by a V-shaped area of low land. Farther on, the river 
occupies a somewhat tortuous channel bordered by narrow alluvial lands 
and then flows along the eastern border of a broad delta area. The 
cutting of the Pansipit Valley has probably been concomitant with the 
elevation of the land. Some writers have expressed the belief that Taal 
Lake was once an arm of the sea and that it contains brackish water and 
a fauna and flora which is in part of marine origin. The water is 
slightly brackish when tasted but there is no evidence that it has been 
salty during recent times because of an admixture of sea water. Its 
character has probably been temporarily changed during eruptions of 
Taal. 

Origin of Bay Lake. — There has been a tendency on the part of some 
writers to consider Bay Lake as having originated through the subsidence 
of a volcano, or to have been formed as a crater. Dana discovered ao 
evidence that the lake corresponds to a single crater, but he considered 
Talim Island as a volcanic center, and Bay Island as consisting of the 
lavas of another vent. As has already been noted, Hochstetter has 
suggested that the part of the lake called the rinconada probably marks 
the position of a subsided crater, but later authors have not concurred 
with his views. Jagor has described Bay Lake, but he has propounded 
no well-defined theory of its origin, although one may infer that he 
considered it to have been at one time an arm of the sea. He found 
a deposit of marine mollusks (principally Tapes virgineus and Cerithium 
moniUferum) 3 meters above the level of the lake, near the point of 
Jalajala Peninsula. He considered these as indicating an elevation of 
the land. Eoth states in his notes that these species were determined 
by Martens and that they belong to the present fauna. 

Von Drasche had no doubt but that the lake was once an arm of the 
sea similar to Manila Bay at the present time, and that it was separated 
by an eruption of Taal, and has since become fresh water. Becker calls 
attention to the fact that Bay Lake is very shallow; the Coast and 



104 ADAMS. 

Geodetic chart shows 6 meters at a point west of Talim Island as 
the deepest sounding. There are but a few places where a boat can 
reach the land. The height of the lake above sea level is stated on the 
Coast and Geodetic chart to be from 0.7 to 1.3 meters. From inquiry 
it was learned that the level of the lake varies as much as 2 meters 
between its lowest and highest stages. The highest stages have been 
produced by exceptionally heavy rains during the wet seasons of certain 
years. It is stated in a footnote in Abella^s report on Mount Maquiling 
that the original site of the town of Bay, which is the oldest of the towns 
founded on the shores of the lake/.wajs. tQ. the west of the mouth of 
Bay Kiver, and is now covered by water. Becker, referring to this 
statement, says that the change is seemingly due to slight earth move- 
ments. Bay is now located on the same river further inland. Martinez 
de Zuniga states that a town called Tabuco, which, according to official 
records existed on the western shore of the lake in 1603, is now under 
water; the settlement having been moved to the present village of 
Cabuyao. Becker also notes that the small island of Sunuli near 
Los Bancs is now united to the mainland. It is probable that the 
name Sunuli refers to Malilimbas Point which contains the crater lake 
called Laguna de los Caimanes. It is described by some as an island, 
by others as a peninsula, depending upon the stage of the water in the 
lake at the time of the observation. The moving of the towns above 
mentioned from their sites may have been brought about by the flooding 
of the lowlands on which they were situated, and possibly the cutting 
of the shore by changes in the channels of the rivers entering the lake 
near them. The writer had the misfortune to spend a night on board 
a • small steamer which was i-un on the mud at the mouth of Bay 
Kiver in order to escape the fury of a typhoon which caused high waves 
in the shallow lake. N'ear the point where the primitive settlement of 
Bay is supposed to have been situated, there are now some houses, and 
during the storm they were easily accessible in hancas. Such a position 
would manifestly be insecure for building a town, and it is not im- 
probable that the original settlement was moved to escape the disaster 
which might be wrought by similar high water. 

Becker states that the lake basin is merely a portion of the great 
plain of Luzon separated from the main area by a slight undulation of 
the surface. The escarpment which is found lying to the west of the 
alluvial lands bordering the lake, has already been explained as a fault 
scarp which has receded westward by erosion. The escarpment descends 
gradually from Mount Gonzales to elevations of about 40 meters and is 
cut through by the valley of the Pasig Eiver to the east of Manila. It 
is probable that this ridge is a part of a fault block which has been 
raised more than the country occupied by the lake, and the elevation 
may have been gradual so that the cutting of the channel of the Pasig 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 105 

River was accomplished as the land emerged. This theory of the origin 
of Laguna de Bay is somewhat similar to one proposed by other writers, 
who have considered the lake once to have been an arm of the sea, but 
it supposes a differential movement of the land to the west instead of 
the formation of a long bar, or spit, of water-laid tuff which upon being 
elevated, cut off the lal^e from Manila Bay. 

Returning to the evidence presented by the finding of marine fossils 
on the shores of the lake, the writer wishes to question the fact of the 
shells having been deposited in marine banks. While studying Jalajala 
Peninsula, and inquiring concerning deposits of shells, he was told of 
a place situated well up on the flank qf the mountain and near the trail 
passing from over the peninsula, where numerous shells could be found. 
The locality was visited and the place was seen to bear many indications 
of having been the site of human habitation. The shells were exposed 
on the surface or slightly buried in the soil, and occurring with them was 
found the bone of a carabao. x\t the present time there are some houses 
near by which are occupied during a part of the year when rice is 
cultivated on the hillsides. It seems altogether probable that the ac- 
cumulation of shells is in the nature of a kitchen-midden. 

In the first part of this article reference is made to a statement by 
^ Smith that he found marine-cut terraces on the east face of a limestone 
ridge in Binangonan Peninsula, and shells of Crassatellites lying on one 
of these terraces. The writer, upon examining the locality, could see 
no evidence of marine-cut terraces, but he did find some shells lying in 
the trail, and it was his opinion that they had been dropped there by 
travelers. It would be very interesting to find beds containing shells 
on the shores of Laguna de Bay and their presence would be entirely in 
accorda;ice with the probable history of the lake; but until authentic 
cases of the discovery of fossil shells are reported it would be well to 
rely on other evidences to prove the elevation of the region. The occur- 
rence of deposits of water-laid tuff at considerable elevations near the 
borders of the lake at other localities than the western shore is a 
sufficiently definite proof of the emergence of the land bordering the 
lake basin. 

MINERAL RESOURCES. 



Lanatin iron deposits. — In his Ligera Resena de 2a Mineria de las 
Islas Filipinas, Abella says under the heading "iron^^: "The deposits 
in the present district of Morong were in reality the first to be exploited 
and smelted, the object being to manufacture munitions of war.'^ 

In discussing the Angat iron deposits, McCaskey quotes the above statement, 
and in his report publishes two analyses of iron >yhich he found in the recordB 
of the Spanish Inspeccion de Minas. Both of these samples were probably from 
93217 4 



106 ADAMS. 

the same locality. The samples designated as from the Lanatin River, town 
of Bosoboso, Province of Morong, showed 66.08 per cent metallic iron; the other 
from Santa Inez, town of Bosoboso, Province of Morong, showed 53.41 per cent 
metallic iron. The reference to Bosoboso probably accounts for the subsequent 
statement by McCaskey in his Fifth Annual Report that there are deposits of 
iron in Bosoboso. 

In visiting the Lanatin deposits the writer journeyed by the way of 
Bosoboso, past the deserted settlements of San Jose, San Isidro, and 
Santa Inez. At Santa Inez there are small bowlders of iron ore in the 
river, and the remains of the abutments of a suspension bridge are in 
part constructed of bowlders of iron ore. Bowlders of iron ore, some of 
which are from 2 to 3 meters in diameter are encountered about one 
hour's walk up the river in the bed of the stream. The mountain to the 
west of the river was evidently the source of these masses. The lower 
slope of the mountain was ascended along the bed of a stream which 
empties into the river just above the bowlders. The country rock 
exposed by erosion is an andesite containing numerous small specks of 
pyrite, and in some places bunches of pyrite were found in sheer zones. 
The larger' masses of pyrite were partially altered to hematite. In 
places there is a small amount of chalcopyrite present and the altera- 
tion has given rise to a coating of the blue and green copper carbonates. 
The copper ores have been prospected lately, but have not been found in - 
encouraging quantities. On the wall of the ravine, a face of rock was 
seen which showed a considerable amount of iron ore, coating and replac- 
ing the country rock. This has somewhat the appearance of a dyke 
running up the mountain, although there is no proof that it is, since 
the dense vegetation obscures the formation excepting in the walls and 
bed of the ravine. ISTear the top of the hill there is an outcrop of iron 
ore. The summit of the hill is capped by a heavy bed of limestone such 
as is frequently met with in the eastern cordillera. In descending, 
exposures of a metamorphosed fine-grained clastic rock were seen in 
the bed of the ravine to the south of the one which was followed in 
ascending. This rock contains specks of pyrite, but no bowlders of 
hematite were seen. A simple and sufficient explanation of the origin 
of the iron ore is that it has been derived from the pyrite which is found 
disseminated in the country rock and occurring as masses in the sheer 
zones. It is probable that the mineralization is a result of contact 
phenomena resulting from the intrusion of the andesite in the sedi- 
mentary formation. 

The amount of iron ore is sufficient to supply a small furnace operating 
as a local industry, and utilizing charcoal in smelting, but there is at 
present no exposure of an ore body which would warrant the establish- 
ment of a large furnace. 

It should be. remarked here that the prospectors who examined this 
locality report another iron deposit somewhat more promising at a 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 107 

distance of a long day^s march in the rugged and very little known 
country to the northeast. 

Iron deposits near Santa Inez, — A deposit of iron which has never 
been worked occurs a short distance to the north of Sainta Maria, in the 
valley of the Santa Maria Eiver which enters the northeast arm of Bay 
Lake. The position of this deposit is shown on d'Almonte^s map of 
Luzon, and to the south of it there is indicated an outcrop of coal. From 
inquiry it was learned that neither the coal nor the iron are of sufficient 
importance to warrant special investigation. 

COPPER. 

• 

Prospects in the Loboo Mountains, — Some prespecting for copper was 
carried on in the northeast part of the Loboo Mountains on the head 
waters of the Calbasahan Eiver. The locality was visited by Smith in 
1905. He described the occurrence of fissure veins which are in two 
sets, the main ones striking in a direction north 70° west, and the others 
north 45° to 57° west. The country rock is a diorite with more or less 
gneissic phases. The principal copper ores found were black oxide of 
copper and bomite with some carbonates, but occuring in small pockets 
and thin veins. Inasmuch as the prospecting did not develop any work- 
able deposits, they were abandoned. 

Prospects on the Lanatin River, — In describing the iron deposits on 
the Lanatin Eiver, reference has already been made to the occurrence of 
copper stains on the iron ores, and occasional small pockets of copper 
ores. The prospecting failed to show an encouraging amount of ore. 

GOLD. 

Many reports have been received concerning the presence of placer 
gold in the rivers which have their sources in the eastern cordillera. to 
the north of Bay Lake and in places some platinum is found with the 
gold. For example, the Mariquina Eiver near Manila carries a little 
fine gold, and prospectors have panned a good number of colors in the 
Novaliches Eiver and in streams further to the north. There is little 
reason to suppose that any of these placer deposits, even if locally work- 
able, will prove important. They have been investigated many times. 

Just to the north of San Isidrb on the trail to Santa Inez, there are 
a number of pits dug near the trail in an area of low land bordering a 
small stream. This locality has been prospected within the last few 
years and some gold found, but not enough to warrant systematic work. 
San Isidro is now within the watershed of Manila water supply reserva- 
tion and further work at the locality is prohibited. However, there is 
. little reason to believe that anyone would care to make further attempts 
at the place. The following explanation of the occurrence of placer 
gold at. the localities described above is offered with some hesitancy. 
It has been shown, in discussing the geology of the eastern cordillera, 



108 ADAMS. 

that dioritic rocks are present and that the iron deposits at Lanatin are 
related to them. At the Lanatin deposits the principal ore is pyrite. 
Ass.ays show that the ore carries some gold. Pieces of rock which came 
from the excavations near San Isidro showed mineralization, and on 
fresh surfaces, well disseminated pyrite. The processes of erosion and 
rock decay have probably liberated the gold associated with the pyrite and 
given rise to the placer deposits. In addition to the pyrite as a source of 
gold there may be some gold-bearing quartz veins in the eastern cordillera, 
but in so far as now known the veins of the region are not well min- 
eralized and no mining claims have been staked on quartz ledges. 



The presence of thin beds of (^oal in the eastern cordillera has held 
out the hope to many prospectors that thicker beds might be found. 
In the belt of sedimentaries associated with the limestones which pass 
Antipolo and the waterworks gorge near Montalban, beds of coal a few 
inches thick have been prospected from time to time within the last 
few years and some work has been done with the hope of discovering 
thicker beds, but without encouraging results. Ickis records the fact 
that coal was found in the belt of sedimentaries which he crossed between 
Infanta and Tanay, but he states that it is an inferior lignite. Further 
south in the same belt of sedimentaries coal has been found near Ati- 
monan. On D'Almonte's map it is indicated as occurring on the Pag- 
bilao Islands east of Lucena, but no reports as to its thickness have 
been received in the division of mines, and there is little reason to 
believe that it is of economic importance since, although well situated, 
it is not' mined. 

Bulacan coal field. — McCaskey while studying the iron deposits in 
the eastern cordillera near Angat found some beds of lignite and received 
information concerning the occurrence of similar deposits near his field 
of work. Kecently, the division of mines has received a sample of coal 
from the valley of the river which passes Norzagaray, and which is 
sometimes called the Matictic. The coal seam is said to have a workable 
thickness. It is exposed in the bed of a stream about 18 kilometers 
from Norzagaray, and beyond that point it is found outcropping in some 
hills. No investigation of these coal beds has been made by the division 
of mines since the information concerning their existence was received 
after the field work for this report was completed. 

GAS. 

In drilling wells for artesian water in the central plains region, gas 
was encountered at Santo Tomas and at Santa Rita, a barrio of Manalin. 
The amount at these places was sufficient to burn with a flame about 
3 meters high when it was first encountered, 'but after a few days the 
pressure diminished. The depth at which the gas was struck at Santo 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 109 

Tomas was about 38 meters. At Santa Rita it was encountered at a 
slightly shallower depth which was not definitely reported. The forma- 
tion is a recent littoral deposit containing beds of shells and much 
organic matter buried in silts. The gas probably has its origin in the 
decomposition of the organic matter, and there is little probability that 
the amount will be sufficient to prove of commercial value since flows 
of gas encountered at shallow depths in similar formations are usually 
soon exhausted. 

In the well which was drilled at Bay in the area of alluvial deposits 
bordering Bay Lake, an intermittent flow of gas was encountered at a 
depth of 69 meters. No record of this well has been received by the 
division of mines, but it is presumed that the gas at Bay has an origin 
similar to that of Santo Tomas and Manalin, with the difference that 
at the latter place the deposits were formed on the border of the lake 
while in the former they are brackish water deposits belonging to the 
river delta area bordering Manila Bay. 



Guadalupe and Meycauayan stone. — The principal building stone 
used in Manila in former years has been the water-laid volcanic tuff ob- 
tained along the banks of the Pasig. It is a soft stone which can readily 
be cut with axes and bolos, and the facility of transportation on the 
Pasig to Manila has rendered it a cheap structural material. A similar 
stone reputed to be of a somewhat better quality is obtained to the east 
of Meycauayan, and it has also been used in Manila. Many quarries 
have been opened within the area of the tuff deposits, but, with the ex- 
ception of the Guadalupe stone from the Pasig River banks and the Mey- 
cauayan, none of the products have received trade names. The tuff has 
low compression and tensile strength and when used in large buildings 
requires the construction of a very thick wall. The fortifications built 
by the Spaniards around the Walled City of Manila are of this stone 
and it was used in public buildings and the numerous churches in the 
capital, and had been shipped by water transportation to many towns 
in the provinces. It is fast falling into disuse except for low walls and 
fences. 

Sisiman andesite. — A large commercial quarry at Sisiman, which lies 
to the east of Mariveles on the north shore of the entrance to Manila Bay, 
was opened for obtaining stone for building the breakwater and harbor 
impovements at Manila. Stone from this locality had been used during 
Spanish times as a building stone and for paving blocks in Manila. 
A good sample of its use as a building stone is found in the Spanish 
Bank building, where it was employed principally in columns and in 
trimming. It has a soft, gray appearance when weathered. Its princi- 
pal defect is its tendency to scale which is exhibited in some blocks. 



1 10 . ADAMS. 

The Sisiman quarry was developed in a small hill which in its central 
portion exhibits a massive columnar structure such as is found in the 
volcanic neck and stock. This structure was favorable to the quarrying 
of large stones which were needed in building the breakwater. A crush- 
ing plant was established for supplying crushed stone and a considerable 
quantity has been used in concrete construction in Manila. 

Basalt. — During the Spanish regime a quarry was opened on a small 
peninsula near the town of Binangonan and some paving blocks were 
cut at localities near by. The quarry had too small a face to warrant 
development on a commercial scale, and, moreover, the rock was variable 
in texture. Some stone was obtained from the locality when the im- 
provement of the streets of Manila was begun by the Americans, but 
in a short time they transferred quarrying operations to Malagi Island, 
which was then used as a prison. 

Certain textural varieties of this basalt are employed for making small 
stone mills such as the natives use for grinding and hulling rice. 

The city of Manila opened a large quarry at Subay in the northeast 
part of Talim Island, and this has been the principal source of nearly 
all of the stone used in macadamizing Manila streets. There the basalt 
occurs in the form of a flow, but it is of variable texture, portions of it 
being vesicular and somewhat scoriaceous. This has rendered it difficult 
to obtain an even product, and the city requested the division of mines 
to look for a more suitable location and a better stone. 

Oabhro. — The Spanish authorities established a quarry southeast of 
Angono for the purpose of obtaining stone for riprapping a part of the 
Pasig Eiver banks near Pasig and building a short breakwater for Ma- 
nila Harbor. The records show that they installed steam drills and a 
tramway to the border of the lake and built a stone pier to facilitate 
the loading of scows by dumping from the tram cars. All the equipment 
was sold and the quarry abandoned before the American occupation. 
The stone pier remains to mark the termination of the tramway and by 
following inland, the quarry may readily be found. The rock is a brec- 
ciated gabbro, suitable fo;r rough stone or crushed stone for concrete 
work, but it is not well suited for macadamizing roads. A new quarry 
site has been selected by me to the northeast of Angono., where the 
gabbro is of even texture. The stone at this place is probably the best 
which can be obtained for use as road material in Manila, and the city 
is making arrangements to open a quarry there and abandon the one now 
in operation at Talim Island. 

In addition to the quarries above mentioned, considerable stone is 
broken up by hand on Talim Island and Binangonan Peninsula. It is 
sold for macadamizing roads near the border of Bay Lake. Formerly a 
quarry was operated at Los Banos by the Army when the road from 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. Ill 

Calamba to Los Baiios was constructed. The rock, which is an andesite, 
is exposed in a high face, and the quality was found to be satisfactory. 

Limestone. — Limestone from east of Montalban has been cut and po- 
lished for the base of columns in some of the churches in Manila and has 
been used for floors, sidewalks, and for paving patios. 



A large amount of gravel is dredged from the Pasig and Mariquina 
Rivers for use in concrete construction, surfacing roads and filling low 
areas. It is of fair quality and makes a good substitute for crushed^ 
stone in concrete work, although it is not so evenly graded, and will not 
stand so great a compression stress. 

Beach gravels and bowlders have been transported from the mouth of 
the Pinamucan River, on Batangas Bay, to the town of Batangas, where 
they were crushed and used in constructing the macadamized roads 
leading to Bauan and San Jose. The bowlders are andesite and have 
suffered considerable disintegration by the action of water, so that the 
crushed stone from them is not as good as would be obtained from a 
quarry of the same rock. 

SAND. 

Sand for use in Manila has been principally obtained from the Pasig 
River by diving and dipping it up in baskets and by dredging. During 
the field work which was done for this report, a large deposit of sand 
was found in the Orani River, which is navigable for launches. It is 
of superior quality and some of it is now being used in Manila. 

Sand has also been obtained from Sangley Point near Cavite, but it is 
of inferior quality. In building the fortifications near the entrance to 
Manila Bay, beach sand consisting largely of fragments of shells has 
been used. There are many places where sand is available, and in build- 
ing concrete bridges and public buildings it is customary to use a local 
supply. The sources above mentioned are the only ones from which 
large quantities have been obtained. 

CLAY AND CLAY PBODUCTS. 

A fair quality of kaolin has been obtained for many years from 
deposits situated on the lower slopes of Mount Maquiling, near Los 
Banos. It is used in making whitewash. It has evidently been formed 
by the disintegration of rocks through solfataric action, since there are 
small solfataras and hot springs near the deposits. A similar kaolin ha^ 
also been worked near Matiquio on the east side of Jalajala Peninsula 
and in a small way to the east of Nasugbu. Within the tuff area there 
are some deposits of fine-grained clays interstratified with the tuff beds 
and occurring in more or less lenticular masses. One of these which has 



112 ADAMS. 

received some investigation is situated near Bocaue. Alluvial clays are 
found in many places and are employed in burning ordinary red pottery 
and soft brick. The pottery industry is carried on principally along the 
Pasig River and in the eastern part of the delta of the Pampanga Eiver. 
The usual product is an unglazed red ware, which is sometimes given 
a wash of red ocher before it is burned, but at San Pedro Macati semi- 
vitrified wares are made. A porcelain factory was established at Manila 
on the banks of the Pasig at a place called Mandaloyan^ by Don Enrique 
Zobel. The foreman in charge of the work was a Filipino who had 
.learned to make pottery during a residence in Japan. The clays which 
were employed were obtained from Los Baiios, Bocaue and Nasugbu. A 
fair grade of porcelain and glazed pottery was produced. In so far as 
the writer has been able to learn, this was the first attempt to establish 
the porcelain industry in the Philippine Islands. It did not meet with 
success, owing to the many difficulties which were encountered in obtain- 
ing materials of suitable .character and of uniform quality. Enough 
was done to demonstrate that there \b a future for the industry. The 
factory is now dismantled. The Bureau of Education has taken some 
steps to teach the making of pottery and stoneware in trade schools, and 
at the present time experimental work is being carried on. 



Common salt is made by the evaporation of sea water at Paranaqu-e 
and Batangas. The brine is usually filtered through a bed of gravel and 
evaporated in a tank having a lime mortar floor. No refined salt is 
produced in the Islands. 

LIME. 

Lime is burned from coralline limestones and corals obtained from 
reefs on the beaches of Tayabas, Batangas, and Balayan Bays and near 
Nasugbu. At Malabon seashells are used. Small limekilns are situated 
on the Pagbilao Islands and near Tayabas and Lucban. Limestone, 
obtained near by, is used. The only important source for limestone is 
the exposure in Binangonan Peninsula. The lime is burned by piling 
it over wood and the product is slaked before being sent to market. In 
fact, quicklime can not usually be obtained in the Islands unless it is 
by special request. One reason for this seems to be the use of water 
transportation and the loading of the lime into hancas and cascoes which 
do not suitably protect it from water, and in case of the entrance of water 
would cause trouble by the heating of the lime when slaking. Small 
quantities of lime are sometimes burned for the manufacture of sugar 
at interior points, near the sugar mills, and for this purpose shells or 
coral are commonly used. 



GEOLOGIC RECONNAISSANCE OF SOUTHWESTERN LUZON. 113 

BIBLIOGRAPHY. 

Abella y Casariego, Don Enrique. El monte Maquilin (Filipinas) y bus 

azuf rales. Madrid, Tello. (ISSS), 1-15. 
Abella and others. Estudio descriptivo de algunos manantiales minerales de 

Filipinas, etc. Manila, Chofr6 y Compa. (1883), 1-50. 
Becker, George F. Report on the Geology of the Philippine Islands. Annual 

Report, U. S. Geol. Survey (1899-1900), Part III, 487-625. 
Centeno, Jos6. 

Memoria sobre los • temblores de tierra ocurridos en Julio de 1880 en la 
Isla de Luz6n. Madrid, Tello (1885), 1-91. 

Estudio geologico del Volcan Taal. Madrid, Tello (1885), 1-53. 
Hochstetter, Dr. Ferdinand. Schreiben an Alexander von Humboldt. Sit- 
zunpsber. Ahad. d. Wiss. (1859), 36, 121-141. This is a report on the 
volcanoes visited during his cruise in the Novara, Luzon, pp. 130-138. 
ICKis, H. M. . 

Report upon building stone near Manila. Sixth Annual Report of the 
Chief of the Mining Bureau, Manila, P. /., (1905), Exhibit E. 

A Geological Reconnaissance, from Infanta, Tayabas, to Tanay, Rizal. 
This Journ., Sec. A, (1909), 4, 483. 
Jagor, F. Reisen in den Philippinen. Berlin, Wiedmannsche Buchhandlung. 

(1873), 16, 381. 
McCaskey, H. D. Report on a geological reconnaissance of the iron region of 

Angat, Bulacan. Bull, No. 3, Mining Bureau, Manila, P. I. 
Riciithofen, F. Von. Vorkommen der Nummulitenformation in den Philip- 
pinen. Ztschr. d. deutschen geol. Ges., (1862), 14, 357-360. 
Roth, Justus. Ueber die geologische Beschaffenheit der Philippinen. In Jagor's 

Reisen in den Philippinen, 333-354. 
Smith, Warren D. 

Orbitoides from the Binangonan Limestone. TJ0S Journ. (1906), 1, 203. 
Preliminary Geological Reconnaissance of the Loboo Mountains of Ba- 
tangas Province. This Journ. (1906), 1, 617. 

Contributions to the Physiography of the Philippine Islands. IV The 
Country Between Subig and Mount Pinatubo. This Journ., Sec. A, (1909), 
4, 19-23. 



ILLUSTRATIONS. 



Plate I. 

Sketch showing the geologic mapping of southwestern Luzon by Von Drasche. 

Plate II. 

Fig. 1. Crater of Taal Volcano seen from the east. 

2. Crater of Taal Volcano seen from the southeast, the 1904 crater in the 
left foreground. 

Plate III. 
Fig. 1. The 1904 crater blowing out steam and gases July 4, Red Lake in fore- 
ground. Photograph by Dean C. Worcester. 

2. The 1904 crater erupting stones. July, 1904. Photograph by Dean C. 

Worcester. 

3. Panorama of Taal crater. New crater formed in 1904 shown to the 

left. 

Plate IV. 

Fig. 1. The 1904 crater, showing erupted mud. 
2. The 1904 crater quiet. 

Plate V. 

Fig. 1. Erosion on outer southern slope of the crater of Taal. 

2. Erosion on eastern inner slope of Taal, showing bedding in crater rim. 

Plate VI. 

Fig. 1. Napayong and Bubuing Islands in Taal Lake, as seen from Li pa Point. 
2. The cone Binintiang Malaqui on the northwest point of Taal Island. 

Plate VII. 

Fig. 1. Marine conglomerates, small island near Corregidor. 
2. Agglomerates, rapids in Pagsanjan Gorge. 

Plate VIII. 

Taal Volcano as seen from Bafiaderos when the steam rises in the morning. 

Plate IX. 
Maquiling Mountain as seen from the railway station at Santo Tomas. 

Plate X. 
Gorge east of Montalban, site of Manila waterworks dam. 

Plate XI. 

Alluvial lands, east border of Bay Lake. 

115 



116 ADAMS. 

Plate XII. 

Fig. 1. Antipolo Falls, on railway to Antipolo. «^ 
2. Botocan Falls, east of Majayjay. 

Plate XIII. 

Fig. 1. First falls, Pagsanjan Gorge. 
2. Upper falls, Pagsanjan Gorge. 

Text Figures. 

Fig. 1. Index map showing physiographic regions of Luzon. 

2. Sketch map showing probable land areas before the emergence of the 

central plain and the plains of the southwestern volcanic region. 

3. Shark's teeth [Squalidae?) found in a railw^ay cut in water-laid tuff 

on the crest of the ridge between Bay Lake and Manila Bay. 

4. Mammalian tooth (cf. antelopes of the Siwalik Pliocene of India) found 

at a depth of between 81 and 85 meters in drilling a well at Pasig. 

5. General geologic section, Infanta to Tanay, Ickis. 

6. Sketch map of the geology and topography of Mount Maquiling and the 

surrounding country, as mapped by Abella. 

7. Sketch showing the extinct cinder cone Cerro de la Mesa, and outer 

crater rim as seen looking south from Calamba railway station. 

8. Sketch of the tufaceous area of Taal Volcano as mapped by Centeno. 

9. Sketch showing the relations of Tagay-tay ridge and the upland on the 

northeast border of Taal Lake as seen from Lipa Point. 
10. Profile of the gorge of the Mariquina River at the head works of the 
Manila water supply, showing river terraces cut in limestone. 

Maps. 

1. Geologic reconnaissance map of southwestern Luzon. 

2. Topographic map of Taal Island. 



Adams : Reconnaissance of Southwestern Luzon.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




Fig. 1. CRATER OF TAAL VOLGANO SEEN FROM THE EAST. 




Fig. 2. CRATER OF TAAL VOLCANO SEEN FROM THE SOUTHEAST, THE 1904 CRATER IN THE LEFT FOREGROUND. 

PLATE II. 



Adams : Reconnaissance of Southwestern Luzon.] 



[Phil. Journ. Sci., Vol. V, No. 2. 





Fig. 1. THE 1904 CRATER BLOWING OUT STEAM AND GASES JULY 4, RED LAKE IN FOREGROUND. 



Fig. 2. THE 1904 CRATER ERUPTING STONES. JULY, 1904. 









■V;JP^ 



m^. 



■1bt. 



Fig. 3. PANORAMA OF TAAL CRATER. NEW CRATER FORMED IN 1904 SHOWN TO THE LEFT. 

P U AT E III. 



Adams : Reconnaissance of Southwestern Luzon.] [Phil, journ. Sci., Vol. V, No. 2. 





-4 


.ii^H^^i^^^B 



im&^^r^ 



m^ 



Fig. 1. EROSION ON OUTER SOUTHERN SLOPE OF THE CRATER OF TAAL. 




Fig. 2. EROSION ON EASTERN INNER SLOPE OF TAAL, SHOWING BEDDING IN CRATER RIM. 

PLATE V. 

* 3 S • 



Adams : Reconnaissance of Southwestern Luzon.] [Phil. Journ. Sci., Vol. V, No. 2. 



H^^^^^^^^S^^ 




1 


F 












^^gfSt: ■:,;:::,■■ • -.. -■ ^ 


•^ 


^^K^-' "' .. . ' ^^ i 


^^"■^^^^^ 






-. ■ _ ... ' ^ 


^p#«**-^^^^^2^^^ 


f^—rz., r>agf/*JftSSw^ 


^'=^,^.sm'f^S'::::VK:,k£^j^:.A 


.^J-agi-rr^--^''*^—^ 


i 


s& 


^^^qgag^^^lg^-^^ 



Fig. 1. NAPAYONG AND BUBUING ISLANDS IN TAAL LAKE, AS SEEN FROM LIRA POINT. 




FiQ. 2. THE CONE BININTIANG MALAQUI ON THE NORTHWEST POINT OF TAAL ISLAND. 



Adams: Reconnaissance of Southwestern Luzon] 



[Phil. Journ. Sci., Vol. V, No. 2. 




Fig 1. MARINE CONGLOMERATES, SMALL ISLAND NEAR CORREGIDOR. 




Fig. 2. AGGLOMERATES, RAPIDS IN PAGSANJAN GORGE. 
F» L- AT E VII. 




X 

H J 



> 







< 

z 
o 

I- 
< 

co 

> 
< 



< X 



H 
z 

D 
O 

z 



< 



o 

> 



Oh 




< 
Z 

o 

Q 
h 



o 



p 

< 




> 



o 

> 




o 

O 




o 



O 

-J 

o 

Q. 

t- 

Z 
< 



o 




o 




Oversized 
Foldout 



Oversized 
Foldout 



CONCRETE CONSTRUCTION IN MANILA AND THE 
PHILIPPINE ISLANDS. 



By W. C. Reibling. 
{From the Chemical Laboratory ^ Bureau of Science, Manila^ P. I.) 



INTRODUCTION. 



Careful, systematic inspection of the sand, gravel and stone, as well 
as of the cement used in all concrete construction, is necessary to secure 
satisfactory permanent results. The aggregates available for concrete 
work in the Philippine Islands vary greatly in quality. Many of the 
sands and gravels which have been used for this purpose are of poor 
quality, and their nature together with that of other available material 
should be thoroughly investigated, since until such information is secured, 
reliable and economic practice in concrete construction can not be assured. 
It is therefore desirable to establish a comprehensive, reliable and 
practical working standard to be embodied in building specifications 
' for projects involving concrete construction, such specifications to be 
based on the results obtained by testing the available materials both 
in the laboratory and in structures. 

CONCRETE CONSTRUCTION. 

Cement of the value of 1,384,202 pesos was imported in the Philippine 
Islands during the fiscal year 1909. These figures represent only a 
small fraction of the total for the concrete constructed, as they do not 
include the cost of the aggregate used and construction expenses. An 
industry involving such an outlay should be fostered and given careful 
scientific assistance. 

Portland cement concrete is eminently fitted to be used as a building 
and structural material in the Tropics and its increasing use in the 
future in the Archipelago can be foretold by its history in other parts 
of the world. In America, the rise in importance of concrete during 
the past ten years has been phenomenal ; in fact, Portland cement manu- 
facture now ranks second of nonmetallic industries in point. of value- 
New ends to which it may be applied with advantage are being discovered 
continually, and new machinery and a better knowledge of the value and 
chiaracteristics of the raw and finished products are reducing its cost. 

117 



118 REIBLING. 

Tropical climatic conditions enhance its value^ for in our climate it is 
never subjected to low temperatures. It is absolutely impervious to the 
ravages of white ants. The local high temperature, heavy rains, and 
alternate humid and dry atmosphere, although rapidly destructive to 
iron and steel and their protecting coatings, are actually beneficial to 
good concrete, adding to its strength, hardness, and durability. 

Properly* constructed concrete will resist earthquake shocks and it 
can be made fire and water proof. It protects iron or steel rods, beams 
or girders from rusting. It is sanitary and clean. The materials can all 
be shipped in small parcels no matter how massive the desired structure, 
thus eliminating many of the difficulties of inferior or limited trans- 
portation facilities. 

However, these influences are advantageous only to good concrete. For 
instance, a bad mixture resulting in a permeable concrete will not 
protect the reinforcement from atmospheric influences, and the iron 
work may rapidly rust away. Cement expands when wet and contracts 
on drying. Therefore, the concrete exposed to sun and rain should he 
homogeneous and its mortar not too rich, otherwise surface cracking 
will result. It must also be remembered that the same raw material 
and identical process of manufacture may result in a 25 per cent difference 
in the ultimate strength of the exposed concrete if made before, during, 
or after the rainy season. Under the present conditions a high factor 
of safety is our only recourse to assure permanent structures. 

The same conditions existed in the ITnited States only a few years 
ago. Concrete construction was taking tremendous strides, but the , 
demand for the more economical use of cement was heard from all sides. 
More reliable construction, better workmanship, more practical building 
laws, and a better knowledge of the value of the natural resources of 
sand, gravel and stone was demanded. As a result, builders, engineers, 
architects, chemists and cement testers carried on experimental work and 
the Government aided the inquiry. The Structure-Material Testing 
Laboratory at St. Louis which is devoted entirely to testing concrete 
material has an annual appropriation of $100,000. Similar work is 
necessary in the Philippine Islands, and fortunately it can be undertaken 
without excessive extra expense, as we already have a fully equipped 
cement-testing laboratory at the Bureau of Science. 

THE SAND USED IN CONCRETE CONSTRUCTION. 

The importance of thoroughly testing sand and gravel has been fully 
set forth in the literature on concrete construction. 

"Upon large or important structures it pays from an economic standpoint to 
make very thorough studies of the materials of the aggregates and their relative 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 119 

proportions. This fact has been seriously overlooked in the past and thousands 
of dollars have sometimes been wasted on single jobs by neglecting laboratory 
studies, or by errors in theory." ^ 

The experience of Spackman ^ and Lesley,^ described in their report 
read before the Eleventh Annual Convention of the American Society 
for Testing Materials, is also of special interest. 

These authors found that cement which would harden with one sand, would 
not do so with another, and that great variations of the strength of cement mortar 
were produced by washing the sand. These variations are shown in the following 
table. 

Comparative test of sand ''A" washed and unwashed, Ottawa sand, and normal 

cement. 



Age. 



7 days-_ 
28 days . 



"A" un- 
washed. 


"A" 
washed. 


Ottawa 
sand. 


Pounds. 
21 
130 


Pounds. 
172 

228 


Pounds. 
250 
438 



However, Spackman and Lesley state that — 

''such extreme cases are not frequent and from the engineering standpoint are 
less dangerous than where the lack of strength due to the sand is less marked. 
When the failure to harden is complete, the effect is so obvious as to insure the 
taking out of the defective material; but where the failure to harden is only 
partial, it may not be discovered during construction, in which event the latent 
weakness in the structure may in time of unusual stress cause failure and 
consequent grave disaster." 

In another report* we find the following: 

"Sand, or the fine aggregate shall be suitable siliceous material passing the 
one-fourth inch mesh sieve, and containing not over 10 per cent of clean, un- 
objectionable material passing the 100 sieve. A marked difference will be found 
in the value of different sands for use in cement mortar. This is influenced 
by the form, size, relative roughness of the surface of the sand grains, and the 
impurities, if any, contained. Only a clean, sharp, gritty sand, graduated in 
size from fine to coarse, and free from impurities can be depended upon for best 
results. Soil, earth, clay, and fine 'dead' sand are injurious to the mortar, and 
at times extremely dangerous; and they also materially retard the hardening 
of the cement. An unknown or doubtful sand should be carefully tested before 
use to determine its value as a mortar ingredient." 

^Taylor and Thompson. Concrete, Plain and Reinforced. New York. (1907), 
183. 

^M., Am. Soc. Testing Materials. 
"Assoc. Am. Soc. C. E. 

*Eduard S. Lanied. Cement Age (1907), 4, 2. 
93217 5 



120 



REIBLING. 



The absolute dependence of the strength of mortar upon the quality 
of the sand is strikingly illustrated in the history of standard sands used 
for cement testing, and the experiences of this laboratory serve to 
illustrate this fact. 

The committee on Philippine cement specifications adopted as a 
standard sand the natural Tarlac Eiver sand, screened through 20-mesh 
to the inch sieve and held on a 30-mesh sieve. With one exception the 
committee saw no reason why the use of this material should not give 
as satisfactory and reliable results as standard Ottawa or quartz sand. 
Tarlac sand was used during many months and was thoroughly tried by 
every method of mixing and various percentages of water; but erratic, 
variable and comparatively low results were always obtained. Even 
when rescreened, the different shipments gave different results; yet to 
all appearances it is a sharp, clean sand, principally containing quartz 
with but a very small percentage of hornblende and weathered soft 
material. A standard cement tested according to the three specifications 
for cement testing then in use gave the following results." 

Tests of a standard cement with Tarlac and Ottataa sand and crushed quartz. 



Specifications. 


Kind of sand. 


Void 

in 
sand. 

Percent. 
36.2 
47.7 
44.7 


^-6 

II 

boo 

SI 


Water. 


Tensile strength. 


Relative 

compressive 

strength. 


7 
days. 


28 
days. 


3 
months. 


28 
days. 


3 
months. 


Am. Soc 


Ottawa 

Crushed quartz. 
Tarlac- 


4,183 
3,445 
3,6ai 


Per cent. 

9.5 

12.5 

»10.7 

21 


305 
287 
220 

606 


364 
327 
264 

635 


395 
330 
302 

646 


1,633 
1,360 
1,240 

3,460 


1,788 
1,620 
1,370 

4.009 


U. S. Army 

Philippine Gov- 
ernment. 
Am. Soc 


Neat cement 











" other percentages of water and methods of mixing gave lower results. 

Tarlac sand was finally condemned as a standard for cement testing, 
and Ottawa sand adopted with very satisfactory results. 

When nearly pure, carefully screened and washed sands differ so 
much in mortar efficiency, what can be expected of commercial materials 
in which the impurities and granularmetric composition introduce still 
other factors influencing their suitability for concrete construction? 

In Manila, sand from the Pasig River, which flows through the city, 
is almost invariably used for concrete construction. Although this 
material is fairly clean and free from impurities, it is the product of 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 121 

the decomposition and abrasion of weathered, poor-grade gravel. It 
contains but little quartz and a considerable quantity of fine "dead" 
sand and shells. Its low percentage of voids is its one good feature. 

Dr. George I. Adams, of the division of mines, Bureau of Science, 
has located several extensive sand beds within practicable distances of 
Manila. The most promising of these is that from the Orani River, 
which enters the northwest part of Manila Bay. 

These two sands have been under test for some time in this laboratory. 
However, it soon became manifest that the commercial products varied 
so much that it was impossible to report on the relative efficiencies of 
the material now being obtained if these be considered independently 
of granularmetric analysis. The best product from the Pasig River 
is probably a sand obtained from washed gravel screenings containing 
a large percentage of pea gravel. This sand could be screened on the 
one-fourth inch mesh. The best grade of Orani sand contains no small 
pebbles and will all pass through the 0.1-inch mesh. Ordinarily the 
least efficient grains in the Pasig sand are the fine ones which predo- 
minate in the so-called ''hanca" sand dipped up by native divers. The 
objectionable material in tl^e Orani sand is the small, soft pebbles which 
are formed on the surface of sand bars exposed at low tide. Sand 
dredged from the channel of the stream can be obtained free from these. 
Accordingly, the specifications for obtaining the best quality of the 
two sands will be quite different. 

Some results leading to the above conclusions have been obtained; 
more comprehensive tests are under way and it is also proposed to 
secure information by testing the aggregates used in actual construction. 
The latter course is necessary because laboratory tests, to be of most 
value, should be conducted under conditions as similar as possible to 
those of actual practice. Spackman ^ found that a concrete which 
hardened at the laboratory failed to do so in actual construction, 
because the stone in the rotary mixer ground the hydrated clay in the 
aggregate to a powder and the paste so produced prevented the cement 
from hardening. 

Another essential requirement to secure proper results is the develop- 
ment of standard methods of testing the materials. We have such for 
iron and steel, cement, stone, and timber, but not for sand and gravel. 
Specifications should be theoretical enough to insure uniformity and 
yet practical enough to meet the conditions of actual construction work. 
The development of strength in concrete is very elusive and dependent 

^ hoc. cit. 



122 REIBLING. 

upon small variations during manufacture and the subsequent exposure. 
Whenever possible, samples of concrete should be taken at the construc- 
tion work, preserved and tested according to specific instructions given 
t9 the engineers and compared with results obtained at the laboratory. 
Also, it would be advisable, with no great outlay of funds as compared 
with the results to be obtained, to have each district engineer examine 
the sand and gravel beds of his territory and send a few barrels of the 
most promising material to the laboratory for investigation. A record 
of short and long time tests, together with other information as to the 
location, quantity, shipping facilities, labor conditions, etc., would cer- 
tainly be of much value for future use. Construction could- be most 
economically specified, and contractors could bid more accurately. The 
demand for this information and its economic value will subsequently 
be illustrated briefly by a few references to work already accomplished 
in this laboratory with the apparatus at our disposal. 

THE GRAVEL AND STONE USED IN CONCRETE CONSTRUCTION. 

The main bulk of concrete is composed of gravel or crushed stone. 
The more of the latter materials which can be used per unit volume, 
the cheaper will be the concrete. The voids in this aggregate are 
filled, and the individual stones are bound together by the mortar. The 
strength of the concrete depends primarily upon: First, the strength 
of the mortar; second, the efficiency of the gravel or stone; and third, 
the density of the concrete. 

In order properly to specify the relative volumes of cement, sand, and 
gravel for concrete it is necessary to know enough about the raw materials 
to ■f'^rmulate the best and most economical mixture suitable for the 
purpose. The use of more than enough mortar to fill the voids and 
cover the gravel adds to the cost and subtracts from the strength. 
Even in a well-proportioned concrete the ultimate strength is often 
limited by the texture or strength of the coarse aggregate. 

EXPERIMENTAL V^-ORK ON GRAVEL AND SAND. 

Pasig River gravel, so much used in Manila, has often proved to be 
of inferior quality. J^Iuch of it is so weathered and decomposed that 
it cracks under the force of a very light blow. Its low efficiency is shown 
in the following results obtained on 6 inch cubes constructed according 
to the method described in Bulletin ISTo. 329 of the United States Geo- 
logical Survey: 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 123 



Compression strength of concrete obtained with Pasig sand and gravel. 
SAND NOT SCREE5NED. 



Proportion by 
volume. 



1:2:4. 
1:2:5. 
1:2:6- 
1:2:6" 
1 : 2 : 6» 



Relative volume, in cubic 
inches. 



Ce- 
nent. 


Sand. 


Gravel. 


54 


108 


216 


43 


86 


216 


36 


72 


216 


36 


72 


216 


36 


72 


2ie 



Water. 



50.0 
41.3 
32.9 
35.3 
40.2 



Total strength 

of 6-lnch cubes, 

in pounds. 



At first 
crack. 



Ulti- 
mate 
strength 



44,280 
46,770 
44,490 
32,000 
34, 740 



53,200 
52,450 
53, 440 
42,870 
39,000 



Compression 
strength, in 
pounds, per 
square inch. 



At first 
crack. 



Ulti- 
mate 
strength 



1,230 
1,300 
1,236 

889 
965 



1,480 
1,457 
1,484 
1,191 
1,083 



Age, in 
days. 



In 

moist 

air. 



In 
open 
air. 



SAND SCREENED THROUGH 10-ME8H SIEVE. 



1:2; 
1:2; 
1 :2 



4 


54 
43 
36 


108 


5 


86 


6 


72 







216 
216 
216 



54.3 
52. 4 



34,600 
22,210 
22,080 



39,200 
33,230 
26,600 



960 
617 
613 



1,090 
923 
739 



■ The effect of gauging with an excess of water is shown in theee two results. Care 
should be exercised not to overdo the "puddling" method. 

When these cubes were crushed, about 40 per cent of gravel broke; 1:2:4, 
1:2:5, and 1:2:6 mixtures gave almost the same crushing strength, indicating tha,t 
the ultimate strength was due to the mortar which was much stronger than the 
gravel. The contract specified a 1:2:4 mixture. The voids in the gravel were 
only 37 per cent, and a 1:2:5 mixture gave a better concrete both in theory and 
in practice. 

The sand contained 20.6 per cent of fine gravel remaining on the lO-mesh 
sieve, 40 per cent of which was fine gravel retained by the 4-mesh. Accordingly 
a 1 : 2 mixture produced a stronger mortar than it would if the sand had contained 
no gravel. As this mortar was sufficient to fill the voids in the gravel, which 
were lowered by the fine gravel from the sand, a stronger concrete resulted. 

It will be difficult to find a substitute for Pasig gravel because of 
its cheapness. However^ gravel has been dredged from the Mariquina 
Eiver which enters the Pasig at a point about 10 miles from its moutli 
and about 8 miles above Manila, which has proved far superior to the 
usual Pasig aggregate since concrete made with it withstood a pressure 
of over 2,000 pounds per square inch and yet showed few broken stones. 

The crushed stone at present used by the city for macadamizing 
streets is not satisfactory for concrete construction, because of its frac- 
ture and texture and the considerable amount of quarry dirt which it 
contains. However, it seems very possible that crushed stone from the 



124 



REIBLING. 



quarry at Sisiman^ near Mariveles across Manila Bay, or from the old 
Spanish quarries near Angono on Bay Lake, and more especially the 
materials located by George I. Adams, of the division of mines, Bureau 
of Science, just north of Angono (which is the best rock for road material 
thus far found near Manila) may be found so efficient as to prove the 
most economical material in the end. The stone from Sisiman with- 
stands a pressure of 1,115 tons per square foot. By using leaner mix- 
tures of properly crushed and graded stone or less massive concrete, it 
may be able to compete in cost with reliable concrete made with Pasig 
gravel. Then, too, important work often demands the best concrete 
obtainable, regardless of expense. 

Sisiman rock unfortunately has a very undesirable, splintery or spally 
fracture which reduces its concrete efficiency. However, if uniformly 
graded it produces a good aggregate for concrete as the following test 

will show: 

Sisiman rock concretes 



Dimensions, in 
inches. 


Propor- 
tions, by 
volume. 


Age. 


Total strength, 
in pounds. 


strength, in 
pounds, per 
square inch. 


In 

moist 

air. 


In 
air. 


First 
crack. 


Ulti- 
mate. 


First 
crack. 


Ulti- 
mate. 


6 X 6 X 6 


1:2:4 


Days. 

7 


Days. 
21 


47, 745 


75, 770 


1, 326 


2, 105 



Mark on cubes. 



" Beach sand was used. The stone was all broken to pass 1.5 inch circular openings 
and the fine material passing J inch mesh sieve was removed. 

Comparative tests between 1:2:5 mixtures of good Mariquina gravel 

and of broken basaltic rock, used with the same cement and sand, showed 

that the rock, despite its larger percentage of voids, gave the more efficient 

aggregate. 

Mariquina gravel versus basaltic rock concrete. 



Dimensions of 
cubes, in inches. 


Propor- 
tions, by 
volume. 


Age. 


Total strength, 
in pounds. 


Strength, in 
pounds, per 
square inch. 


Mark on cubes. 


In 

moist 

air. 


In 
air. 


First 
crack. 


Ulti- 
mate. 


First 
crack. 


Ulti- 
mate. 


6 X 6 X 6 

6 X 5 X 6 


|--_1 : 2 : 5 


Weeks. 
3 


Weeks. 
9 




1 


__1,550 
__2, 124 


2,026 
2,380 


Mariquina gravel. 
Basaltic stone. 




\ 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 125 

Many gravels it seel for concrete structures in other parts of the Phil- 
ippine Islands have proved themselves to be worse than the poorest 
material from the Pasig River. From time to time concrete cubes taken 
from the mixtures used for construction work have been submitted to 
this laboratory for valuation. In many cases the concrete proved to 
be of poor quality and almost without exception it was evident that the 
weakness of the cubes was due to other causes than that of the quantity 
or quality of the cement, which, it must be understood, had previously 
been tested and accepted. 

In many cases the gravel was so poor as to crack and crumble under 
light loads. For instance, two cubes of the concrete used for Bridge 
No. 14^ P. & A. Road, Tayabas, were well proportioned and molded 
concrete. '^K^^ concrete was richer and harder than "B,^' and yet both 
cubes possessed the same poor strength. iVs about 50 per cent of the 
gravel cracked and crumbled under the light load recorded, it is evident 
that the strength of the concrete was limited by the inferior quality of 
the gravel. 

The results obtained were the follo\ving: 

Concrete cubes from Bridge A'o. IJf, P. cC- A. Road, Tayahas. 



Dimensions, in inches. 


Propor- 
tions, by 
volume. 


Age. 


Total strength, 
in pounds. 


Total strength, 
in pounds, per 
square inch. 


Mark on cubes. 




First 
crack. 


Ulti- 
mate. 


First 
crack. 

331 
438 


Ulti- 
mate. 

619. 
612 


4X4X4 _ _ i 


1:2:4 
1:3:6 


Days. 
28 
28 


5, 305 
7,015 


9,900 
9,800 


Class "A". 
Class "B". 





The results obtained with cubes taken from the concrete used for 
Carcar Bridge^ Province of Cebu, project E-16, are especially instructive. 
The rich class ^^A^' concrete was stored in water two weeks and then 
exposed in the open air for the remainder of the sixty days. • Con- 
sequently the cement had ample opportunity to harden. The mortar 
in each case was rich and durable and in each the strength obtained 
varied directly as the quality of the gravel used. 



126 



REIBLING. 

Concrete cubes from Carccur Bridge, Gehu. 



Mark on cubes. 


Age.ir 
days. 


Dimen- 
sions, in 
inches. 


Initial 
crack. 


Ultimate 
strength. 


Strength, in 

pounds, per square 

inch. 


Remarks. 


















First 
crack. 


Ultimate, j 








/ 49,730 


65,000 


1,381 


1, 806 1 The mortar is rich. 

There were a 

f number of bro- 




i 








ken stones, espe- 


Class "A" con- 


i 








! cially the large 


crete, South 


i 








ones. This cube 


Arch ring, 


1 








i contains 3 stones 


C a r c a r 










about 3 inches in 


B ri dge, 


1 








diameter. 


Province of 


i 


55, 500 


63,800 


1,542 


1,772 This cube is the 


Cebu, proj- 












1 same as the above 


ect R-16 


61 


6X6X6 


i 






except that it 

does not contain 

! large stones of 3 




j 


! 






inches in di- 




j 










ameter. 


1 




35,300 


48,600 


980 


1,350 


The mortar is good. 


' 






i 






1 Aboutone-halfof 














, the stones were 














broken. The 














! stones used were 














rather soft. 


i 
1 






37,750 


39,650 


1,048 


1,101 Do. 








i 44,700 


63,150 


1,242 


1,754 The mortar is good. 
i Small sized 
1 stones were used. 
j This cube was 
1 well proportion- 
1 ed. A small per- 
; centage of the 
j stones were 




i 








j broken. 



Cubes stored two weeks in water and then in open air. 

The same gravel was crushed with a weaker mortar (Class ''B'' con- 
crete) and gave results as recorded below. The extremely low strength 
at the "first crack" illustrates the greatest danger in the use of poor 
gravel. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 127 
Concrete cubes from Carcar Bridge^ Cehu. 













Strength, in 
pounds, per square 




Mark on cubes. 


Age, 

in 
days. 


Dimen- 
sions, in 
inches. 


Initial 
crack. 


Ultimate, 
strength. 


inch. 


Remarks. 


First 
crack. 


Ultimate. 


Class "B" con- 
















crete, Spand- 






10,600 


43,600 


294 


1,211 


Good mixture, fair 


rel Wall. Car- 














mortar. The lar- 


car Bridge, 














ger stones were 


project R-16_ 


35 


6X6X6 










mostly broken. 
Gravel poor. 








28,650 


54,630 


796 


1,517 


Do. 1 








(•) 


35,300 




980 


First crack at about 












8,000. 


Do 


37 


6X6X6 


21,300 


28,680 


592 


794 


The large stones 




were broken. 
















Sand seems to be 
















the chief fault, as 
















it contains shells 
















and dirt. 


Do 


47 


6X6X6 


9,250 


32,800 


257 


911 


Voids in sand not 


! 


filled up. The i 
















mortar was weak. , 


Do 

i 
1 


37 


6X6X6 


17,300 


28,000 


480 


778 


Large percentage 
of the stones was 
broken, fair mor- 
tar. 








(^) 


36,000 




1,000 


First crack at about ' 








i 






6,000 pounds. ! 


Do 


47 


6X6X6 




i 
i 






Cube received 
with 2 corners 


















broken off. 


! 
i 






16,000 


34,000 


444 


944 


The mortar as well 
as the gravel was 
poor. 








( 17,800 


36,800 


480 


1,022 


The majority of the . 


Do 


49 


6X6X6 


, 








stones were 
broken. 




i 






I (b) 


36,400 




1,011 


First crack at about 












6,000 pounds. 
















Cube received 










i 






with 2 corners 






i 




1 






broken off. 



a Less than 8,000. 



bLess than 6,000. 



All cubes were submerged two weeks in water. 



128 



REIBLING. 



Plate I is a photograph of the second to the last cube recorded in the 
above table. Each arrow points to the center of the exposed surfaces of 
crushed or cracked stones. As can be observed, the mortar between these 
stones is almost perfectly sound. 

A more specific description of the stones designated by the arrows is 
as follows : 



Arrow No. 



3 

4 

7 !_..__. 

8 

9 

11 

13 

17 

5, 6, 10, 12, 14, 15, 16, 18. 



Dimensions 

of exposed 

surface, in 

inches. 



1.5X0.6 
1.5 X 1.5 

1. 25 X 0. 6 
1.0X0. 75 
1.5X1.0 
3.0X1.0 
1.2X0.5 
2 X 1. 25 
1. 3 X 0. 6 

1.25X1 



Remarks. 



I 



Large piece of wood charred on the outside. i 

Large round stone of soft, yellow substance which is about ' 

as durable as dry clay. Stone was completely crushed. ' 

Do. I 

Do. 

Do. 

Do. 
Harder than above. One end cracked off as illustrated. 
Large stone, cracked in two places. 
Large stone, split through the center. 

Do. 
Small cracked and crushed stones. 



The floor concrete from the Naga Kiver Bridge also contained poor 
gravel, and again showed poor strength despite good mortar and proper 
proportioning. 

Concrete cubes from ^aga River Bridge, Province of Gehu.^ 



Dimensions, in inches. 


Propor- 
tions, by 
volume. 


Age. 


Total strength, 
in pounds. 


Total strength, 

in pounds, per 

square inch. 


Mark on cubes. 


First 
crack. 


Ulti- 
mate. 


First 
crack. 


Ulti- 
mate. 


6X6X6 ._ 


1:2:4 


Days. 
30 
30 
30 
30 
29 
29 


12, 255 
11,750 
20, 105 
14, 055 
16, 685 
8,460 


21, 285 
20, 820 
28,000 
25, 940 
18, 525 
11,260 


340 
326 
558 
390 
463 
235 


591 
578 
778 
720 
515 
213 


P-1. Floor. 
P-2. Floor. 
P-3. Floor. 
P-4. Floor. 
P-5. Floor. 
P-6. Floor. 





"The mixture Is good and the mortar hard, but the gravel is fine and of poor quality. 
Much of it broke through weathered material. 

At other times good enough gravel has been used ; but poor sand, weak 
mortar or improper proportioning has caused poor concrete. A few 
examples will serve to illustrate this. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 129 

Concrete cubes from Naga River Bridge, Province of CehuA 



Dimensions, in inches. 



5. 96X 5. 95X 5. 95 . 



Propor- 
tions, by 
volume. 



Age. 



Days. 
28 



Total strength, 
in pounds. 



First 
crack. 



(9, 570 
14, 740 



Ulti- 
mate. 



18,505 
925 



Strength, in 
pounds, per 
square inch. 



First 
crack. 



272 
185 



Ulti- 
mate. 



527 
271 



Mark on cubes. 



Class A. 
Class B. 



»The appearance of these cubes showed that the mixture employed was an improper 
one, as there was not enough mortar present to All the voids properly. The mortar was 
also weak and easily broken with the fingers, and had an unnatural, brown color. It is 
possible that impurities such as clay prevented the hardening of the cement. 

Concrete cubes tested as per ''Request No. 68328'' gave erratic results 
which were attributed to excess of sand and to the poor grading of the 
gravel. Very little of the gravel was found to be cracked in any of 
the crushed cubes, but its granularmetric composition w^as extraordinarily 
poor. With the exception of several stones 2.5 to 3 inches in diameter, 
the gravel was of a uniform size. Fine gravel to reduce the voids was 
lacking. 

The results obtained were the following: 

(Compression tests of concrete cubes per Request No. 68328. 



Dimensions, in inches. 



5^X6^X6__ ___- 

6X6i X6 

6X6X6 

6X5^X6 

52X6iX6 

5iX6X6 



Class. 



"A" concrete 

__-_do 

do 

-__.do 

."B" concrete 

do- 

"C" concrete 



Age. 


Total strength. 

in pounds, per 

square inch. 


Total strength, 
in pounds, per 
square inch. 




First 
crack. 


Ulti- 
mate. 


First 
crack. 


Ulti- 
mate. 


Days. 










37 


42, 970 


54,295 


1,194 


1,508 


37 


35,000 


52, 570 


972 


1,460 


37 


54,450 


60,060 


1,512 


1,670 


43 


26,250 


36,210 


730 


1,006 


44 


14,500 


37,500 


403 


1,042 


44 


27, 420 


34,310 


762 


953 


23 


8,470 


34,220 


235 


945 



Mark on cubes. 



"A" floor. 

Do. 

Do. 
"A" pier. 
"B" pier. 

Do. 
"C" No. 2 pier. 



When a good aggregate and a proper mixture had been employed, the 
concrete gave much better results. A few samples are here recorded 
for comparative purposes. 



130 



RBIBLING. 

Concrete cubes from Bayaoas River Bridge.* 



Dimensions, in inches. 


Class. 


Age. 


Total strength, 
in pounds, per 
square inch. 


Strength, in 
pounds per 
square inch. 


Mark on cubes 


First Ulti- 
crack. mate. 


First 
crack. 


Ulti- 
mate. 








Days. 










6X6.1X5.9 


"A" concrete 


29 


23,980 


62,980 


655 


1,720 


Al. 


6. 1 X 6. 1 X 6. 9 


do 


29 


19,180 


66,090 


516 


1,775 


A 2. 


6.1X6.1 X5.9 


do 


29 


29,850 


68, 575 


802 


1,840 


A3. 


6.1X5.9X 6 


"B" concrete 


26 


41,990 


63, 495 


1,164 


1,762 


Bl. 


6.0X6.1X5.9 


do _- 


26 


31,790 


63,770 


866 


1,743 


B2. 


6. X 6. 1 X 5. 9 


do 


" 


29,550 


60,265 


807 


1,666 


B3. 



* Gravel was very good, and very few broken stones were found in the crushed concrete. 

Compression tests. 



Mark on cubes. 


Age in 
days. 


Dimen- 
sions, in 
inches. 


Initial 
crack. 


Ultimate 
strength. 


Strent 
pound 
square 

First 
crack. 


rth, in 
J inch. 

Ulti- 
mate. 


Propor- 
tion. 


Class "A" concrete from con- 
crete used in R. C. piles, Pan- 
tal Bridge, LL-3, Pangasinan, 
October 16. 1909. 


28 


6X6X6 


42,900 

39,000 

" 60,000 

44,000 


51,000 
50, 100 
62,000 
52, 930 


1,192 
1,083 
1,666 
1,222 


1,417 
1,392 
1,722 
1,470 


1:2:4 
1:2:4 
1:2:4 
1:2:4 


Average 


46, 425 


54,000 


1,291 


1,500 









The above cubes were well proportioned, the mortar as well as the concrete 
contained few voids. Very few broken stones were found. The aggregate is 
composed of a standard cement, Santa Barbara sand, and Sual Well gravel. The 
cubes were kept constantly moist under wet cloth. 

Plate II is a photograph of the third concrete cube recorded in the above 
table. About 70 stones were exposed on the surface of the crushed cube. 
Of these, only two small pebbles were cracked. The strength of the 
stone was greater than that of the mortar. The photograph shows cracks 
in the mortar in many places. 

Such analytical results and conclusions lead to the assumption that 
by synthetic reasoning we could foretell the value and strength of a 
concrete by a careful study of the aggregate from which it is to be made. 
For instance, we would not expect that soft, coral rock concrete would 
possess good strength despite a proper mixture and a rich strong mortar, 
as we would assume that the soft stone would limit the strength of the 
concrete to a very low figure. We would also assume that if harder 
coral rock were substituted for the softer material the resulting concrete 
would be stronger. Accordingly, soft and hard coral 1:2:4 concrete 
was manufactured ; and when tested gave the following average results : 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 131 

Compression tests of coral rock concrete. 



Crushed stone. 


Propor- 
tions 

by vol- 
ume. 


Age. 


Average total 
strengths, in 
pounds, per 
square inch. 


In 
moist 
air. 


In air. 


First 
crack. 


Ultimate. 


Soft coral rock 

Hard coral rock 

Hard coral rock 


1:2:4 
1:2:4 
1 :2:4 


Days. 

7 
7 
7 


Days. 
22 

22 


731 

782 

1,029 


852 
1,005 
1,644 









In every instance the majority of the stones in the concrete were 
broken or crushed. However, the harder coral rock concrete was stronger 
at 7 days than the other at 29 days. 

Three sands and four stones were available for a certain concrete 
structure. They were all sent to this laboratory for inspection and test- 
ing. The report is interesting inasmuch as it illustrates the great 
variability in the efficiency of different available materials and the ac- 
curacy and value of preliminary inspection. 

Briefly summarized the report was as follows : 

Sand tests ( preliminary report ) . 



Sand. 



Beach _ 



No. 1- 
No. 2. 



Specific 
grav- 
ity. 



2.62 
2.74 



Voids, 
dry. 



Description. 



Per ct. 
47.3 



45.0 
41.7 



Very tine in grain; 10 per cent "dead sand;" poor granular- 
metric composition. 
Clean, calcareous, and of good granularmetric composition- 
Similar to sand No. 1, but better in all respects i 



Order 

of 
merit. 



Mortar tests {final report). 



Sand. 


Tensile strength, in pounds, per square inch. 


Relative com- 
pression strength 
of mortar cylin- 
ders (7 days in 


1:2 by volume. 


1:3 by volume. 


1 : 3 by weight. 


moist air; 21 
days outside). 


7 days. 

285 
306 


28 days. 

398 
378 


7 days. 


28 days. 


7 days 


28 days. 


1:2 by 
vol- 
ume. 


1:3 by 
volume. 


Beach 


183 
170 
135 


260 
200 
240 


162 
128 
188 
308 


225 
194 
242 
366 


1,788 
2,790 
2,910 


1,550 
1,597 
1,715 


No. 1 


No. 2 


Standard Ottawa 





















Sand No. 2 gives the best mortar results. 



132 



RBIBLING. 

Gravel tests — Preliminary report. 





Specific 
gravity. 

2.56 

2,47 
2.68 
2.62 


Voids 
(per 
cent). 

45.8 

46.3 

37.1 

•44.7 


Ideal mix- 
ture. 


Physical properties. 


Beach gravel 


1:1.73:3.6 

1:2:6 
1:2:5 
1:2:4 


Some stones are durable and hard, 
but the majority are so weathered 
as to disintegrate under slight pres- 
sure. The efficiency of the gravel 
for concrete is poor. 

Fine coral and shell gravel, not 
coarse enough for gravel and too 
coarse for sand. It is unsuitable 
for concrete. 

Durable, hard beach gravel with 
rounded and pitted surfaces. Good 
concrete material if properly 
graded. 

Hard and tough stone, which despite 
its undesirable spally fracture is a 
good concrete aggregate. 


No. 3 gravel 


No. 4 gravel _ 


No. 5 stone (crushed Sisiman) 



n Graded. 

Concrete tests {final report). 

CONCRETE CUBES. 



No. 


Dimen- 
sions, in 
inches. 


Propor- 
tions, by 
volume. 


Age. 


Total strength, 

in pounds, per 

square incn. 


Total strength, 
in pounds, pei' 
square inch. 


Mark on cubes. 


In 

moist 

air. 


In 
air. 


First 
crack. 


Ulti- 
mate. 


First 
crack. 


Ulti- 
mate. 


1 


6X6X6- 


1:2:4 


Days. 

7 


Days. 
21 


20, 800 


41,430 


580 


1,151 


(Beach sand. 
iBeach gravel. 


2 


6X6X6- 


• 1:3:6 


7 


21 


8,000 


28,330 


222 


800 


Do. 


3 


6X6X6- 


•1:1.73:3 


7 


35 


23,840 


56,890 


662 


1,560 


Do. 


4 


6X6X6- 


"1:2:4 


7 


35 


19,560 


4,500 


543 


1,250 


Do. 



» Ideal mixture, •' Soft stone removed. 

Nos. 3 and No. 4 show a majority of the stone broken. Mortar good. 



6X6X6- 

6X6X6- 
6X6X6_ 

6X6X6- 
6X6X6- 

6X6X6- 



1:2:5 

1:2:4 
1:3:6 

1:2:4 
1:3:6 

1:2:6 



21 


33,835 


48, 845 


968 


1,357 


21 


20, 130 


55,500 


560 


1,540 


21 


15,215 


32, 495 


423 


903 


21 


47, 745 


75, 770 


1,326 


2,105 


21 


24,660 


34, 715 


685 


964 


21 


16,820 


25, 580 


468 


710 



No. 4 gravel. 

No. 2 sand. 
(No. 5 stone. 
I No. 1 sand. 

Do. 
(No. 5 stone. 
\no. 2 sand. 
Do. 

No. 3 gravel. 

No, 1 sand. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 133 



No. 1, very few stones broken. No. 2, a considerable number of broken stones. 
No. 3, soft mortar and broken stones. No. 4, some stones broken, mortar fair. 
No. 5, very soft mortar. No. 6, very soft concrete. No. 2 sand gives the highest 
efficiency, as is shown by No. 5 gravel cubes. 

At another time laboratory and field tests did not agree. 1:2:4: 
mixtures when tested after seven days in moist air and twenty-one days 
in air withstood an average pressure of 1,491 pounds per square inch. 
Over 50 per cent of the gravel was cracked. That which cracked was 
andesite containing glassy feldspars in a soft matrix. The 1 :2 :4, 5, and 
6 mixtures again gave practically the same strength. Later, we received 
from the contractors concrete cubes which gave higher results. In these 
cubes all the andesite gravel with glassy feldspars present, cracked, but 
whereas the original gravel sent to us contained about 75 per cent of this 
friable andesite, that in the contractors' cubes contained only about 33 
per cent of this material and some large stones over 2^ inches in di- 
ameter. The cubes had been submerged in water for six days and then 
exposed to the air, thus producing a very strong mortar. The cone 
showed that the surrounding mortar was sound in many places where 
the stone had crushed. 

The facts above mentioned show the necessity of proper representative 
sampling and a uniform method for the treatment of concrete samples 
after they have been gauged. The same concrete preserved under 
different conditions will give variable results. Cubes are sometimes 
received which have apparently not been wetted since gauging and con- 
sequently show poor strength when tested, even though the sand, gravel 
and cement are good and have been mixed properly. A previous inspec- 
tion of material often enables the engineer to improve the quality by 
simple methods. A few examples will serve to illustrate this. 

Sometimes is its advisable thoroughly to wash the material. At other timei^ 
this expense is not necessary. Thus Novaliches sand gave much higher tests 
washed than unwashed, while that used for the Bocaue Bridge contained less than 
3 per cent of dirt and fine material and its removal was not advantageous. 

Novaliches Sand {dirt and fine material 5.0 per cent by weight). 



Sand. 


strength 1:2 mortar, 7 days. 


Tensile. 


Compres- 
sion. 


Per cent 
efficiency 

tensile 
strength. 


Novaliches (washed) 

Novaliches (unwashed) 

standard quartz 


288 
224 
308 


2,485 
2,087 


93.5 
72.7 









134 REIBLING. 

Sand for Bocaue Bridge {dirt and fine material 2.8 per cent hy weight. 



Sand. 


Tensile strength, 7 days. 


1:2 
mortar. 


1:3 
mortar. 


Per cent 
efficiency 
1:3 mortar. 


Bocaue Bridge (washed) 

Bocaue Bridge (unwashed )__ 
Standard quartz 


283 
277 


170 

172 

248 


68.5 
69.3 











Sometimes the graniflarmetric composition of the sand can be much improved 
by screenings from the gravel, and vice versa. Thus, samples of the only available 
sand and gravel for the Trade School building at Malolos, sent to the laboratory, 
proved of good quality in every respect but that of its granular metric composition. 
The sand was very fine, 51.5 per cent passed through the 40- and remained on 
the 50-mesh sieve. The gravel was clean and hard, but contained 40.7 per cent 
of sand passing the 4-mesh sieve. 

It frequently happens that aggregates do not permit recomposition, but in this 
case it was easily accomplished. 

The gravel was screened on a 4-mesh sieve and the coarse sand thus obtained 
was mixed with the fine in the proportion of 3 to 1. The percentage of voids in 
the sand thus produced was reduced from 40.4 to 36.2 per cent. 

Comparative compression tests showed a gain in efficiency of almost 60 per 
cent, and a dense, impermeable concrete was obtained even with a 1:3:6 mixture. 
It was found practicable to follow this method in practice and the materials were 
so used. 

Relative compression strength of Malolos concrete using sand, and a mixture of 
this sand and sand screened from the gravel. 





Mixture 






Relative compression 
strength. 


1 cement, 


3 sand. 


6 gravel 






. 622 


1 cement, 


3 sand mixture, 6 


gravel 


(1 




sand, 


3 sand 


screened 


from 


the 




gravel ) 










992 


1 cement, 


2 sand, 


4 gravel 






1,230 



The examples given, though incomplete and stated briefly, are con- 
sidered sufficient to show the variable nature of available sands and 
gravels in the Philippines and to indicate the value of reliable and 
exact knowledge concerning their characteristics. However, such tests 
are not to be relied upon implicitly, because they were not made under 
working conditions, and the time was too short in which to complete 
them. The material should be tested before the building specifications 
are drawn up, and not after the materials are delivered. In some cases 
tests have been put off until the time for construction had nearly arrived 
and then only a few pounds of the sand and gravel were sent to the 
laboratory for a five or seven day test, which was necessarily of question- 
able value. If this work is worth doing at all, it is worth doing well. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISLANDS. 135 

CONCRETE SPECIFICATIONS. 

The following quotations will serve to demonstrate the weight laid 
upon concrete specifications in other part^ of the world. 

"In framing concrete specifications, the proportions of the constituents should 
be stated so distinctly that there can be no misunderstanding between the engineer 
and the contractor as to the quantities which will be required for the work." ® 

"Engineers have placed the volume of a barrel of packed cement all the way 
from 3.1 to 3.8 cubic feet, corresponding to a variation in the weight of from 
123 to 100 pounds per cubic foot. Loose cement measurements, on the other 
hand, are variously fixed at from 3.8 to 4.5 cubic feet to. the barrel, or 100 to 
84^ pounds per cubic foot. The extreme actual variation is therefore from 3.1 
to 4.5 cubic feet per barrel, or 123 to 84^ pounds per cubic foot. The quality 
of cement should ' invariably be regulated by its weight",® as has been done in 
France and England. 

At this laboratory we have adopted the American Society standard of 
100 pounds of cement being assumed as equal to 1 cubic foot. Sand 
and gravel are measured dry. 

"Without a universal standard there is no means of comparing the concrete 
in different structures or the results of different experiments, and an unscrupulous 
contractor may adopt for his unit the volume of cement very loosely measured, 
and thus produce too lean a concrete." ^ 

He may also fail to reject "cake^^ cement, unless such action is 
specifically designated in the concrete specifications. 

Although the nature of the sand and gravel or stone is as important a 
consideration as that of the cement itself, many local building specifica- 
tions simply state that the sand shall be sharp and clean and that the 
gravel shall all pass a certain sized sieve. There are sometimes no 
definite statements as to the kind of materials, their chemical composi- 
tion, the percentage of voids, the percentage of dirt, gradation of the size 
of the particles, and the strength they should give when mixed with the 
cement to be used. 

In one instance, which has come to my notice, the contract called for a 1:2:4 
mixture. The gravel and sand for the work were piled in two large heaps and 
samples sent to the laboratory for inspection. As already recorded, it was found 
not advisable to use the 1:2:4 mixture specified. The gravel being well graded 
and the percentage of voids correspondingly small, an excess of mortar resulted, and 
this, woi^cing to the top during the "puddling," produced a layer of pure mortar 
and a non-homogenous concrete. A 1:2:5 mixture was then substituted and gave 
better results. 

Though they seldom do so, concrete specifications should insist upon 
sand-free gravel. Sand tends to reduce the percentage of voids in the 
gravel. Ordinarily a low percentage of voids is desirable, but when a 

* Taylor and Thompson, page 217. 

• Loc. cit. 
93217 6 



136 



REIBLING. 



mixture is already specified and the ratio of mortar to gravel already 
efficient, a further reduction of the voids in the gravel by means of sand 
can only result in lowering the quality of the concrete. 

For example, in one instance a 1:2:5 concrete was specified, but the gravel 
being used contained 22.1 per cent sand (fine material passing 0.51 centimeter 
rectangular openings). Therefore, the actual mixture employed was far inferior 
to that intended by the specifications. This is shown by the following figures: 



Gravel. 



Sand. 



As received 

Screened on 4-mesh sieve- 



Voids. 



Relative 
volumes 
2/5 sand 

+ 1 
gravel 



Actual pro- 
portions, by 
volume, of a 
1:2:5 mix 
i ture. 



Per ccjit. Per cent 



22. 1 




29.1 
33.9 



1.25 
1.215 



1:3.1:4.8 
1:2 : 5 



Cement-] 
sand I 
ratio. 



Mortar- j Volume 
gravel ! of excess 
ratio. ! mortar. 



1 : 3.1 ] 
1 :2 



I 



2.86 : 4.8 ! 
2.16:5 



1.41 
0.44 



The ratio of cement to sand was decreased and the mortar strength was 
weakened accordingly. The volume of mortar became so excessive that it 
separated the gravel or flushed to the surface. The volume of concrete per unit of 
cement was increased and therefore the strength of the concrete was decreased. 

Unfortunately, it is impossible to specify a fixed mixture under the 
assumption that the fine material is uniformly distributed in such a way 
as to make adequate allowance for the sand present in gravel or broken 
stone, since subsequent rehandling and weathering tend to concentrate 
the fine material from the coarse. 

Carefully selected, representative samples taken at different times from one 
large gravel pile showed variations in sand contents of 4.3, 8.0, 13.0, 20, and 
24.8 per cent. The gravel taken from the bottom of this pile contained the 
most sand, ana concrete taken from the mixer when it was being used had less 
strength at the end of eighty days than a like mixture of the same cement, sand 
and gravel (except that the sand was removed from the gravel) at only twenty- 
nine days, as the following results will show: 

Pasig sand and gravel concrete. 



Dimensions, in 
inches. 



6X6X6 



Propor- 
tions, by 
volume. 



Ago. 


Total strength, 
in pounds. 


Strength in 
pounds, per 
square inch. 


In 
moist 
air. 


In air. 


First 
crack. 


Ulti- 
mate. 

52,450 
32,705 
44, 125 


First 
crack. 

1,300 
770 
455 


Ulti- 
mate. 

1,457 

908 

1,225 


Days. 
f ^ 

\ ^ 

7 


Days. 
22 

73 

73 


46, 770 
27, 720 
16,390 



Mark on cubes. 


No sand 


in the 


gravel. 




Considerable 


sand in 


gravel.* 


Do. 





* The 80-day cubes showed few broken stones, but the mortar was weak and porous. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISIiANDS. 137 

In like manner, the sand as delivered was not of uniform character, the per- 
centages of gravel in it varied from 5.5 per cent to 38.6 per cent. 

Samples of Mariquina River sand and gravel were used for another structure 
where the specifications called for a 1:3:6 mixture. The sand was screened 
through a sieve with ^ -centimeter rectangular openings, and the gravel washed 
with water thrown over the pile before it went into the mixture. The effect of 
the water was to remove the dirt and make the sand adhere more persistently to 
the gravel. Analysis of the washed gravel gave the following: 





Per cent 


Held on 3. 81 -centimeter openings 


1 


Held on 2.54-centimeter openings 


2.2 


Held on 1.70-centimeter openings 


10.3 


Held on 1.14-centimeter openings 


27.8 


Held on 0.76-centimeter openings 


49.1 


Held on 0.51 -centimeter openings 


63.1 


Held on 0.38-centimeter openings 


77.3 


Held on 0.25-centimeter openings 


87.0 


Through 0.25-centimeter openings 


91.2 



Percentage of voids 



31.5 



This gravel contained 36.4 per cent sand; therefore the 1: 3 mortar relied upon 
by the designers to bind the gravel was weakened to a 1 : 5.18 in actual con- 
struction. 

For similar reasons 1:2:4 concrete made from crusher run Sisiman stone 
showed greater strength when the dust and fine mateial had been removed from 
the stone. The concrete cubes recorded below were made at the Bureau of Naviga- 
tion, division of port works, and tested here. 

Compression tests of 1:2:4 Sisiman stone concrete. 



No. 



Age. 



Days. 
24 
24 

10 
10 



Streneth,in 
pounds, per 
square inch. 



First 
crack. 



1,080 
1,309 



427 
385 



Ulti- 
mate. 



1,319 
1,744 



601 
771 



Remarks. 



Made with Orani sand, a standard cement, and crusher run stone. 

Same as above with all material passing one-half inch screen re- 
moved from stone. 

Made with Pasig sand, a standard cement, and crusher run stone. 

Same as above with all material passing one-half inch screen re- 
moved. 



Because of these results it was assumed that the fine material in this stone 
was detrimental to its value. However, this was not proved. If we measure 
and consider the fine material passing the one-fourth inch screen as sand and 
make allowance for it as such in the mortar mixture, different results may be 
obtained. A 1:1.5:4 mixture may show better results than the best 1:2:4 
mixture recorded and give as great a volume of concrete as with the fine material 
screened out. 



138 



HEIBLING. 



The logical conclusions of the results recorded in this discussion would 
seem to be a modification of concrete specifications in one of two direc- 
tions, either they should state that the concrete must show a certain 
strength in a given time^ or they should state the granularmetric com- 
position of the screened sand and gravel to be used. 

The first method does not insure either the future of the concrete, its 
surface appearance, or its economical use. Its enforcement would im- 
mediately result in higher bidding. The testing would also be expensive 
and might cause much delay. Failure to meet the required tests would 
result in long discussions and eventually prove ruinous to all interests 
involved. 

The second method would insure safe building and future durability. 
The work could be rushed as rapidly as possible, the sieving and screein- 
ing being done to meet the requirements. It would thus be possible to 
calculate the mixture more closely and actually to .reduce the cost of 
construction. 

Sand and gravel should always be screened dry. With durable rotary 
screens the cost should be very low. One screening of the sand and one 
of the gravel would probably be sufficient for most work. This would 
insure fixed ratios in the morter and concrete and it would then be 
necessary to allow only for variations in the percentage of voids. 

Recently, also, another source of danger to concrete construction became 
manifest when the following cubes were tested : 

Concrete cubes from Ahatan River, Cortes, Bohol. 



Dimensions, in 
inches. 



6X6X6_ 



Propor- 
tions. 



Age. 



In 
water. 



Bays. 
11 



In 
air. 



Total strength, 
in pounds. 



First 
! crack. 



Ulti- 
mate. 



Days. I 
12! 



12,590 ! 25,335 

10,955 I 20,395 

10,845 I 26,420 

Average = 



Strength, in 
pounds, per 
square inch. 


First 
crack. 


Ulti- 
mate. 


349 


704 


304 


566 


301 


734 


318 


668 



Mark oii cubes. 



CI -A. Floor. 
Do. 
Do. 
Do. 



Examination of the crushed concrete revealed the fact that the stones 
were covered with green algae, the effect of which is illustrated by Plate 
III, which -is a photograph of the cone resulting from one of these 
crushed cubes. 

Prom this photograph, which shows the exposed sockets with their 
bordering, thin ridges of sound mortar, it is seen how effectually the 
coating of organic matter prevented adhesion between the mortar and 
the gravel. 



CONCRETE CONSTRUCTION IN THE PHILIPPINE ISIANDS. 139 

The stones were submitted to Mr. E. D. Merrill, botanist of the Bureau 
of Science, for miscroscopic examination. His report is as follows : 

Examination of the pebbles shows that they are more or less covered with a 
very thin, greenish or yellowish film, which is formed by minute, tmicellular 
green algse, apparently Protococcus. 

The pebbles were undoubtedly secured from the upper strata of some stream 
bed, for these algse are abundant in streams throughout the Philippines. 

Gravel should never be taken from the surface layers in the stream bed, for 
this organism is always present in greater or less qualities, both in still and 
swift water. The upper few inches should be discarded, and the gravel of lower 
layers used. 

If surface gravel is used, the film of algse should be removed by heating the 
gravel sufficiently to destroy all vegetable matter on the surface of the pebbles. 
It is possible that the exposure of the gravel to the sun and air for a few weeks 
would cause the disappearance of the algse, but this is uncertain, especially in 
wet weather, for these low forms of plant life are very resistant. 

The simplest method of avoiding difficulties in connection with the use of 
gravel that is more or less covered with algae is that noted above : the removal and 
discarding of the surface layers. 

From the experience of Spackman and Lesley,^^ and of S. E. Thomp- 
son,^ ^ we find that vegetable matter has also a markied effect on the 
strength of mortars. It is less readily detected in sand, but when present 
gives the moist silt a greasy nature which interferes with the hardening 
of the cement. In the Philippines and other tropical countries, where 
in rainy seasons, especially, low forms of plant life flourish so vigorously, 
the concrete workers should exert extra precaution against this source of 
danger to the strength of their structures. 

CONCLUSIOK. 

The discussion and data recorded above lead to the following con- 
clusions : 

1. Adequate consideration is not at present given to the fact that the 
nature of the aggregate is fully as important as the quality of the cement 
itself. 

2. It is a mistake to use good Portland cement with poor aggregate. 

3. The efficiency of concrete is limited to that of its most inferior 
constituent. In a normal mixture the strength and durability of con- 
crete can not be insured to exceed that of either its gravel, stone, or 
mortar; and for any given mixture the mortar efficiency depends as 
much upon the nature of the sand as upon the quality of the cement. 

4. Concrete specifications must include standard requirements for 

^* Sands: Their Relation to Mortar and Concrete. Proc. Amer. Soc. for Treat- 
ing. Materials, (1908) 8, 451. 
"/5i(Z., 449. 



140 



BEIBLING. 



sand, gravel, and stone, as well as for cement, in order to guarantee 
satisfactory results. 

5. Concrete specifications, to insure economy as well as efficiency, 
must be based upon a comprehensive knowledge of the relative merits of 
all available material. 

6. To ascertain the relative merits of the available material, a standard 
method of testing must be adopted. 

7. This standard method of testing must embrace practical conditions 
of manipulation and exposure, so regulated as to subject the concrete 
to equal, uniform influences. 

I am indebted to Mr. F. D. Eeyes, assistant in the chemical laboratory 
of this Bureau, for his assistance in the tests of materials which are 
reported in this article. 

Note. — Shortly after this article had been prepared for publication, 
our attention was attracted to the concrete qualities of basaltic rock from 
the Province of La Laguna. This rock is dense, fine-grained, and hard ; 
and it possesses an ideal fracture for concrete aggregate as it crushes 
into cubical fragments. 

The strength of the concrete made with this crushed stone proved 
far superior to any so far tested at this laboratory. The figures obtained 
were as follows : 

Compression tests of concrete made from crushed basaltic rock from La Laguna. 



Dimensions of 
cubes, in inches. 

6X6X6 

6X6X53 

6X6X6— 


Proportions, 
by volume. 

■ 

1 : 2.2 : 5 

1:2.75:6.25 

1:3.3:7.5 


Age. 


Total strengtli 
in pounds. 


In 

moist 

air. 


In air. 


First 
crack. 

133,000 
76,010 
22,450 


Ulti- 
mate. 


Days. 
21 
21 
21 


Days. 
80 
80 
80 


152,500 
96,900 
36,890 



Strength, in 
pounds, per 
square inch. 



First 
crack. 



3,700 

2,203 

623 



Ulti- 
mate. 



4,236 
2,808 
1,025 



Mark on cubes. 



"B" concrete. 
"C" concrete. 
"D" concrete. I 



The stone was crushed so that all passed 1.5-inch circular meshes. Good Pasig 
sand (gravel-screenings) was used. The cement was the same as used in previous 
laboratory work. 



The strength developed by the leanest mixture recorded above shows 
the economic possibilities of good aggregate in concrete structures where 
great strength is not essential. The 1 : 3.3 : 7.5 concrete proved stronger 
than many 1:2:4 mixtures of poor aggregate. 

The ideal, cubical fracture of the crushed stone, and the dense, masonic 
concrete it tends to produce is illustrated by Plate IV, which is a 
photograph of the crushed, 6-inch, 1 : 2.2 : 5 concrete cube which withstood 
a pressure of over 150,000 pounds. 



ILLUSTRATIONS. 



Plate I. Crushed cube of concrete from Carcar Bridge, showing disintegration 
of gravel. 
II. Crushed cube of concrete from Pantal Bridge, Pangasinan, showing 
disintegration of mortar. 

III. Cone from crushed cube of concrete from Abatan River, Cortes, Bohol, 

showing effect of using gravel covered with green algoe. 

IV. Crushed cube of concrete made from first-class aggregate. 

141 



Reibling : Concrete Construction.] 



[Phil. .Tourn. Sci., Vol. V, No. 2. 




CRUSHED CUBE OF CONCRETE, SHOWING DISINTEGRATION OF GRAVEL. 
RL-ATE I- 



Reibling : Concrete Construction.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




CRUSHED CUBE OF CONCRETE, SHOWING CRACKS IN THE MORTAR. THE STRENGTH OF 
THE STONE IS GREATER THAN THAT OF THE MORTAR. 



R U AT E I I 



ReiblinCx : Concrete Construction.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




CONE FROM CRUSHED CUBE OF CONCRETE FROM ABATAN RIVER, CORTES, BOHOL 
SHOWING EFFECT OF USING GRAVEL COVERED WITH GREEN ALG/€. 



PLATE IN 



Reibling: Concrete Constkuction.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




CRUSHED CUBE OF CONCRETE MADE FROM FIRST-CLASS AGGREGATE. 
RUATE IV. 



-•- o • 



PHILIPPINE POTTERY. 



By George I. Adams and Wallace E. Pratt. 
{From the Division of Mines, Bureau of Science, Manila, P. /.) 



The purpose of this report is to give a brief resume of the pottery 
industry in the Philippines and record what has been learned through 
practical experience concerning the clays of the Islands by thoee who 
have attempted to produce better grades of ware than the primitive pot- 
tery made by household industry. Information has been gathered con- 
cerning the methods used and the results obtained at the first kilns 
built for making stoneware and porcelain from white clays. In the 
future these data may not be so easily collected and they are, therefore, 
recorded here for reference. The investigation of the industry at this 
time was made in order to cooperate with the Bureau of Education in 
the establishment of pottery schools as a part of manual training instruc- 
tion. It has afforded an opportunity for making the first collection 
representative of Philippine pottery. 

Primitive pottery. — The natives of the Philippines make crude red 
pottery at many places throughout the Islands. It is estimated that the 
value of the total output in 1909 was 400,000 pesos. The products are, 
for the most part, pots for carrying water and cooking food. Some of 
these utensils are provided with lids. The designs seldom show any 
special merit and the ornamentation is usually confined to lines scratched 
with a stick in the unbumt ware. Frequently, the air-dried pots are 
washed with a ferruginous earth or ocher ground in water in order to 
give them a uniform red color when burned. 

The clay used is generally alluvial. It is mixed with water, stirred 
or worked over to take out all sticks and pebbles, and then is tramped 
with the bare feet to make it plastic. Most of the clays show consider- 
able air shrinkage and to overcome this as much as one-third part of fine 
river sand is sometimes mixed with the prepared clay. 

The pots are turned by hand on a small wooden wheel which. h£«,8 a 
foot resting on a second piece of wood. The wheel is revolved by hand. 
It is not centered on a bearing or axle, but the foot of the wheel is 
lubricated with a little watery clay. In shaping the mouth of the pots a 
piece of cloth wet in clay water is held between the thumb and finger. 

143 



144 ADAMS AND PRATT. 

After the pots are formed they are allowed to dry until the clay is 
leathery. They are then given a second shaping by holding a smooth 
stone on the inside and beating them with a wooden paddle. This 
renders the clay more dense and serves to overcome the tendency to craok 
while air drying. Sometimes the pots are given a second beating. If 
they are washed with ocher they are usually nibbed with a smooth object 
to cause the ocher to enter the surface of the clay. The operculum of 
some large shell is commonly used in this operation. The air-dried pots 
are then piled on a layer of bamboo sticks or rice straw^ covered with 
more fuel and burned in the open, the burning being completed in about 
twenty minutes. 

The manufacture, of pottery in this way is carried on as a household 
industry. The making of the pots is usually done by women. The 
obtaining of the material and the marketing of the product is done by 
the men. Certain towns and barrios are centers of the industry and in 
passing through them one sees the products displayed for sale in the 
houses and dooryards and hears the beating of the pots as they are being 
formed. The localization of the manufacture seems to depend to some 
extent upon the pressence of suitable clay, but more particularly upon 
the facilities for marketing the product along routes of water trans- 
portation. 

Introduction of kilns. — The use of kilns and some improvements over 
the primitive methods of making pottery seem to have been introduced 
under the Spanish regime in order to meet the demands of household 
articles not found in use by the natives and to supply earthemware 
required in certain industries. The use of kilns permitted a better burn- 
ing of the common red ware and they produced a sufficiently high 
temperature to semivitrify the product. 

Braziers for cooking with charcoal or wood fires are made with three projections 
which will support a round-bottomed pot. Sometimes a small oven forms a part 
of a brazier. If made as a separate piece, the oven can be placed on top of a 
brazier or used as a "Dutch oven" by setting it on a bed of coals and placing fuel 
around it and on the cover. As a substitute for the porous water jars of small 
size, large ones are turned on potters* wheels supported by a vertical axle and 
moved by the foot of the operator or by an assistant. Such jars are burned to 
semivitrification and are used for water jars in households, for collecting the 
various saps used in making native liquprs, and as containers for coconut oil. 
As containers they are sometimes transported on pack animals, but more com- 
monly in the native boats called hancds or cascoes. Conical-shaped receptacles 
with a hole in the bottom called "pilones^' are used in the manufacture of sugar. 
Thd ci*ilde sugar is poured into these molds, where it crystallizes, the molasses 
filtefring but tlirough the hole. The sugar cones or "sugar hats" thus formed 
are sometimes taken from the earthernware molds, but are more commonly 
transported in them. Large and small flowerpots of common red ware are also ' 
made and occasionally bowlshaped receptacles for domestic use. 



PHILIPPINE POTTBEY. 



145 



There are but a few kiln^ in the Philippines. The usual type is a long 
semicylindrical structure built on a grade of about 10 per cent^ having 
a* chimney at the upper end and a door and hearth at the lower (fig. 1). 
There are also openings for firing along the sides of the kiln. They are 
constructed of volcanic tuff and common brick laid up and plastered 
over with mud. The fuel used is wood. A high temperature can not 
be obtained in these kilns and the semivitrified products from some 
of them are due to the low fusing point of the clays used. There is 
considerable shrinkage in burning, and when the clay fuses the sand 
used to prevent cracking comes to the surface, making the ware rough 
and pebbly. 

Attempts to make porcelain. — There are many deposits of white or 
light-colored earths or clays in the Philippines, which are sold in the 




B*io. 1. — Sketch of Philippinb Pottery Kiln. 

markets in the form of balls or cylinders, and used in preparing a kind 
of cold-water paint or whitewash for painting houses. These clays are 
commonly, but improperly, called ''yeso/' which is the Spanish word for 
gypsum. They usually contain a low .percentage of silica, but some 
of them approach true kaolin in chemical composition. 

In 1903 M. Tagawa, a Japanese for many years engaged in business in Manila 
and now proprietor of a store on Plaza Moraga, built a small kiln at Bocaue with 
the idea that the white clay found there could be used in making porcelain. It 
was soon discovered that the clay was not very plastic and burned to a light 
weight, friable biscuit. The kiln was operated for a short time, the output being 
bowls, flowerpots, etc., with metallic oxide glazes on biscuits burned from Bocaue 
clay and common alluvial clay. 

A more serious effort to manufacture porcelain was made by Don Enrique Zobel, 
who built a kiln and installed some machinery near the Roxas hacienda on the bank 
of the Pasig River, west of San Pedro Macati and a short distance beyond the city 
limits of Manila. He became interested in the project through the representations 



146 ADAMS AND PRATT. 

of Sr. Quinto, a Filipino who resided for some time in Japan and investigated 
the manufacture of porcelain there. Experienced Japanese workmen were employed 
and clays from Bocaue, Los Bafios, and Nasugbu and silica from near Laoag in 
Ilocos Norte were shipped to the factory. Some Japanese clays and silica were 
imported. The factory was operated for about one year and a large amount of 
ware waft made and sold, but the enterprise was not successful. It soon became 
evident that a high grade of porcelain clay had not been found and that there was 
a deficiency in silica which could not be readily supplied from Philippine sources. 

The writer has secured irom Yasuziro Kato^ who was employed in 
the factory, three authentic pieces which illustrate the results obtained 
with Philippine materials: 

( 1 ) A yellowish plate ornamented with a fish sketched in blue. The biscuit was 
burned from a mixture of ten parts Los Bafios clay and four parts Bocaue clay. 
The glaze contains "silica" from Laoag. The ware is light weight, not very strong, 
and somewhat pitted. 

(2) A grayish plate ornamented with a fish sketched in blue. The biscuit was 
burned from a mixture of ten parts Los Bafios clay and four parts Batangas 
clay. Laoag "silica" was used in the glaze. The ware is heavy and apparently 
strong, but is somewhat pitted. The biscuit is faintly yellowish. 

(3) *A white bowl (now broken) with a Japanese character for the word "new'' 
written in blue in the bottom. The biscuit was burned from ten parts Los Banos 
clay and two parts Laoag "silica." The glaze was of the same material. This 
ware is the best which was made from Philippine materials. The biscuit is white 
and the glaze white, but not quite so pure as is desirable. 

A better class of ware was manufactured from Los Bafios clay and Japanese 
silica. An example of this product is a plate with a blue transferred Japanese 
design. 

At the Zobel factory some fine porcelain was made from Japanese 
materials and a few articles were decorated in colors and gilt. A great 
variety of articles were manufactured, including small vases such as • 
would readily sell to the trade, and flowerpots, etc., with metallic oxide 
glazes were made from alluvial clay obtained at the factory. It is un- 
fortunate that the factory did not continue in operation, producing such 
ware as it was possible to make from the alluvial clays and finding a 
ready market for their product in Manila. 

Dr. David P. Barrows, formerly Director of the Bureau of Educa- 
tion, advocated the establishing of a pottery school as a part of the in- 
struction in manual training. The services of Kanjiuro Kato, a Japan- 
ese familiar with porcelain manufacture, were secured in 1906 on the 
recommendation of Mr. Hunter Sharp, American consul at Kobe. So 
many difficulties were encountered in the experimental work which he 
carried on that the undertaking was temporarily abandoned. 

Manufacture of glazed pottery. — In 1906, Y. Kato, who had worked 
for Tagawa at Bocaue arid later was in the employ of Zobel, went to 



PHILIPPINE POTTERY. 



147 



Binan and began experimenting with clays from various deposits, includ- 
ing one at Muntinlupa. He built a kiln at San Pedro Tunasan and waa 
joined by K. Kato. 

They use principally alluvial clays, but some of their better ware is made 
from white Calamba clays. Metallic oxide glazes are employed and the product 
includes flowerpots of ornamental design, water jars with faucets and filters, 
vases, jardiniers and pedestals, and lion heads, also statuettes of saints, and 
busts of Rizal sold either as bisques or painted in colors. The output is sold 
principally in Manila. The type of kiln used is shown in the accompanying 
illustration ( fig. 2 ) . 

At Pandacan in Manila, there is a kiln owned by Severino Alberto, 
a Spaniard, which has been in operation for many years. The labor is 
performed by Chinese. 

Alluvial clay is used which burns to a red biscuit. The ware is sometimes 
painted in colors, but usually is given a metallic oxide glaze. Water bottles 
are also made which are given a wash of white clay and then slightly reburned. 
The products of this kiln are principally flower pots, bowls and various earthern 
vessels. The output sells readily in Manila. The ware is not so good as that 
made at San Pedro Tunasan, since no attempt is made to use the better grades 
of clay. 




Fig. 2. — Type of Kiln Introduced by Japanese. 

At San Felipe Neri, Go Kicho, a Chinaman, has a brick kiln and a pottery 
kiln. He uses alluvial clay. Some crude pottery with metallic oxide glaze is 
made in addition to the usual product of red pottery. 

The biscuits which are used for the glazed pottery at the kilns just 
mentioned are made from alluvial clay, but are more dense and much 
stronger than the native ware. This is due largely to a careful washing 
of the clay, subsequent thorough kneading to make it plastic, and burn- 
ing at a higher temperature. The white clays used at San Pedro Tu- 
nasan are employed because they are of better quality than the alluvial 
clays. Being low in silica they do not have the strength requisite for 
stoneware, and containing some iron they exhibit a variable pinkish or 



148 ADAMS AND PRATT. 

yellowish color when burned. Some pieces have been made in which the 
silica has been increased by the addition of Laoag "silica/' 

At San Pedro Tunasan the clay is washed in barrels or tanks. It 
is placed in the first barrel with a large amount of water and is stirred 
with a stick until a considerable percentage is in suspension. A part of 
the water is then dipped from the barrel and poured through a sieve 
having about 60 meshes to the inch, into a second barrel. After a few 
minutes, when the clay is partially settled in the second barrel, the upper 
part of the water is dipped or siphoned back into the first and the 
process repeated. The washed clay which settles as a thin mud is placed 
in the tank between the ovens and the chimney of the kiln and the excess 
of water is evaporated by the heat from the latter when it is fired. At 
Pandacan the clay is washed in pits. In the first pit it is mixed by 
tramping in water which comes to the height of a man's waist. The 
water with the clay in suspension is poured through a sieve into a second 
pit, from which it may flow to either of two large drying floors. The 
process of washing the clay by tramping is not an inviting occupation, 
and stirring the clay in barrels is an awkward method. A log washer 
turned by hand, an arrangement of screens and tanks, and a small pump 
for running water to the washer from the settling tanks would be inex- 
pensive improvements which would save labor and dignify the work. 

Characteristics of white clays, — Some idea of the character of the 
white clays may be gathered from what has been noted concerning the 
results obtained at ZobeFs factory, where the utility of certain of them 
was determined in a practical way. In addition, the chemical composi- 
tion of the clays analyzed in the chemical division of the Bureau of 
Science has been published by Dr. Alvin J. Cox,^ together with some 
physical tests. The writer in the following table has grouped certain 
of these analyses according to localities. An analysis of Laoag silica by 
L. A. Salinger is appended. 

* Occurrence, Composition and Radio-activity of the Clays from Luzon. This 
Journal Sec. A, (1907), 2, 6, 413-435. Laguna Clays, ibid. Sec. A, (1908), 
3, 5, 377-388. 



PHILIPPINE POTTERY. 



149 



Table I.— Analyses of Philippine clays. 









Fluxes. 








o 


Region, province, locality, 
sample number. 


1 


i 

< 


o 
^9 


i 


i 


i 


i 


B 


1 


*■ 


1 


S 

< 


O 


o 

B 
2 


1 
1 


< 


3 


a 
S 

H 


§ 


OS 


Maquiling, Mount, Laguna 




! 






Province: 






















No. 2 Los Banos 


44.15 


36.54 


1.04 


0.15 


0.00 


0.98 


2.17 


1.14 


13.50 


2.64 


No. 3 Vicente Jesus 






















fnine 


43.32 
42.06 


41.48 
32.04 


0.32 
0.33 


0.04 
0.48 


0.59 
0.32 


0.88 
1.17 


1.69 
2.90 


0.00 
1.10 


14.12 
20.42 


0.00 
2.08 


No. 5 Laguna Province 


No. 6 Los Banos 


45.24 


37.31 


1.00 


0.66 


0.00 


1.69 


3.35 


0.98 


12.67 


1.24 


No. 7 Pajo Arroyo 


44.30 


37.28 


1.47 


0.39 


0.42 


1.00 


3.28 


L36 


12.56 


1.60 


No. 12 Calamba 


61.98 


26.22 


0.12 


0.60 


0.72 


0.34 


1.78 


0.00 


10.55 


0.00 


No. 19 Wolfson mine— 


49.95 


31.84 


3.96 


0.86 


0.70 


0.64 


5.66 


0.00 


11.90 


0.00 


Bagong B 1 a 






















Creek 


49.42 


30.45 


1.61 


0.00 


0.21 


0.16 


1.98 


Lll 


11.72 


5.86 


Bagong B 1 a 


Creek 


43.83 
55.99 


3L86 
28.77 


6.86 
0.89 


0.14 
0.18 


0.11 
0.03 


(•) 
0.17 


6.11 
1.27 


0.80 
0.91 


15.04 
11.59 


2.71 
2.42 


Pajo Canon 


PajoCafion 


42.23 


37.32 


1.41 


0.23 


0.07 


0.46 


2.17 


1.00 


15.84 


1.92 


PajoCafion 


43.28 


37.85 


3.39 


0.08 


0.04 


(•) 


3.51 


1.25 


14.20 


0.89 


Point Alipasio— 


43.16 


38.64 


1.19 1 0.09 


0.14 


0.10 


1.52 


1.54 


14.55 1 1.4-2 1 


Nasugbu, Batangas Prov- 






1 








1 ! 1 


ince: 






! 1 












No. 13 Nasugbu 


62.78 


23.85 


L60 


0. 15 1 0. 00 


2.61 


4.36 


0.78 


8.84 


0.00 


No. 20 East of Nasugbu. 


65.18 


19.07 


3.93 


0. 00 0. 28 


L16 


5.27 


0.70 


9.04 


0.00 


No. 35 Near Nasugbu___ 


57.45 


18.08 


8.40 


1.41 0.00 


1.42 


11.26 


0.58 


8.56 


6.08. 


Albay Province: Locality 








! 












not known 


71.16 


16.94 


0.48 


0.57 


0.23 


6.10 


7.38 


0.58 


3.22 


L56 


Matiquio, Jala Jala Penin- 


1 


0.841 4.73 


sula, Laguna Province—. 


61.00 


19.71 j 3.49 


0.34 


0.07 


0.95 


8.18 


5.97 


Malinta, Bulacan Province: 
















No. 30 Tinajeros River, ' 


] i 












alluvial clay 


J 60.24 1 18.73 i 7.19 1 1.78 1 2.06 


1.86 


1 13.04 


0.00 


13.84 


0.00 


San Pedro Macati, R i z a 1 


i 






Province: 1 ' 
















No. 38 Pasig River, 


1 


i 














Mandaloyan, alluvial 






1 . ! 
1 1 












i clay 


. 52.53 


2L01 


8.40 


4.04 • 2.58 I 2.68 


17.70 


0.00 


9.08 


0.00 


Laoag, Ilocos Norte Prov- 






















ince: ' 




















Granulite dike, source l 








1 




j TiO 
1 trace 
\ MnO 
I 0.46 






of "silica" 


. 72.56 


15.13 


2.54 


2.01 


0.95 


1 5.62 


0.00 




. 0.03 





• Small. 



150 ADAMS AND PRATT. 

It will be seen that the clays from the region of Mount Maquiling, 
commonly referred to as the Los Baiios and Calamba clays, are high in 
alumina and relatively low in silica. Some of them approximate the 
composition of kaolinite. The clays from the region of ISTasugbu in 
Batajigas Province are low in alumina and relatively high in silica. It 
will be remembered that a strong ware made at Zobel's kiln was obtained 
by mixing 10 parts of Los Baiios clay and 4 parts of Batangas clay, which 
brought up the percentage of silica, and that the best and strongest ware 
from Philippine materials was obtained by using 10 parts of Los Baiios 
clay and 2 parts of Laoag "silica." 

The clay from Matiquio on the east side of Jalajala Peninsula, Laguna 
Province, is low in alumina but is high in iron. In addition to the 
physical tests reported by Cox, a practical test has been made by burning 
some ware made from it in the kiln at San Pedro Tunasan. It is not 
certain that the clay used in the ware is the same as the sample analysed. 
On burning it showed a peculiar, old-rose color. The practical test 
demonstrated that it is suitable for hard earthenware in which the color 
would be covered by a glaze. 

The analyses of alluvial clays show a rather low percentage of silica 
and a high total of fluxes. They fuse at a low temperature and show 
more shrinkage than is desirable. 

An analysis of a clay from Albay Province, locality not definitely 
known, shows the highest percentage of silica of any of the Philippine 
clays outside of one sample from ISTasugbu, and is inserted for the sake 
of comparison and as an example of silicious clay such as it is desirable 
to find. 

In general it may be stated that the Philippine clays are deficient 
in silica, and, moreover, that the only source of additional silica thus 
far discovered is at Laoag, in Ilocos Norte, far removed from the pottery 
industry which at present centers near Manila. According to Dr. W. D. 
Smith, of the division of mines, this material is obtained from a decom- 
posed granulite dyke.^ It is disintegrated by heating, so that the ma- 
terial which was used in Manila at Zobel's kiln, and to a small extent 
at the bottle factory, was received in the form of a sand. 

Pottery hiln at the Santa Cruz trade school. — Mr. A. G. McLeod, 
superintendent of the schools of Laguna Province, in December, 1909, 
in accordance with the plans of the Director of Education, had con- 
structed under the direction of K. Kato a down-draft kiln with two ovens, 
and began the making of pottery from alluvial clay. It is proposed soon 
to introduce the use of clay from Matiquio which will permit of pro- 
ducing a better ware. This venture promises to meet with success, since 
there is a good local demand for pottery. 

^The Asbestos and Manganese Deposits of Ilocos Norte, with notes on the 
geology of the region. This Journal, Sec. A, (1907), 2, 3, 145-175. 



PHILIPPINE POTTERY. 151 

The estahlishment of a pottery school at Manila. — It is the plan of 
the present Director of Education, Frank E. White, to establish a pottery 
school on the bank of the Pasig River in Manila in connection with 
other courses in manual training now taught at the capital. The services 
of Y. Kato will be available and there is every reason to believe that the 
school will be a success, since the ware of the class which will be made 
is now sold in sufficient amount in Manila to show that the output of 
the kiln will at least pay the running expenses of the school. 

In connection with the school at Manila; it is proposed to carry on 
experiments with clays from the various deposits found throughout the 
Islands with the hope of discovering materials which will make a superior 
ware. The clays from Nasugbu, Los Bancs, and Matiquio, the value 
of which has already been determined, can be economically transported 
to Manila, which is a central point. The Bureau of Science will co- 
operate in so far as is possible in the experimental work. It is probable 
that the Bureau of Education through instruction will be able to dis- 
seminate a knowledge of pottery making throughout the Islands and that 
a considerable improvement in the character of Philippine pottery will 
result from its efforts. 

93217 7 



ILLUSTRATIONS. 



Plate I. 



Fig. 1. Turning pottery on a Chinese wheel. 

2. Finishing a "tenaja." 

3. Transporting "vino" in "tenajas." 

Plate II. 

Fig. 1. Turning pottery on a native wheel. 

2. Beating pots into final shape. 

3. Native pottery market, Ilocos Sur. (Photograph by Dean C. Worcester.) 

Plate III. 

Fig. 1. Pile of pottery, showing method of burning. (Photograph by Dean C. 
Worcester. ) 
2. Pottery peddler in suburbs of Manila. 

Text Figures. 

1. Sketch of Philippine pottery kiln. 

2. Type of kiln introduced by Japanese. 

153 



Adams and Pratt : Philippine Pottery.] 



[Phil. Journ. Scl, Vol. V, No. 2. 





'/'■'■/l/'^ ^ 







Fig. 1. TURNING POTTERY ON A CHINESE WHEEL. 




Fig. 2. FINISHING A TENAJA." 




Fig. 3. TRANSPORTING *'VIN0" IN "tENAJAS.' 
PLATE I. 



• J 



Adams and Pratt : Philippine Pottery.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




Fig. 1. TURNING POTTERY ON A NATIVE WHEEL. 




Fig. 2. BEATING POTS INTO FINAL SHAPE. 




Fig. 3. NATIVE POTTERY MARKET. 
P U AT E II. 



Adams and Pratt : Philippine Pottery.] 



[Phil. Journ. Sci., Vol. V, No. 2. 




Fig. 1. A PILE OF POTTERY, SHOWING METHOD OF BURNING. 




FiQ. 2. POTTERY PEDDLER IN SUBURBS OF MANILA. S « ;••» J"" I'o i* 

E3 I A-t- B- ■ I I » * O 9 



PLATE III 



THE PETROGRAPHY OF SOME IGNEOUS ROCKS OF THE 

PHILIPPINES/ 



By Joseph P. Iddings. 
( United States Geological Survey, Washington, D. G. ) 



PETROGRAPHY. 



The rocks of the Philippine Islands, so far as known, are : 

1. The igneous rocks that form the volcanoes and principal mountain 
ranges, and that consist of massive lavas, both extrusive and intrusive, and 
to a larger extent, perhaps, of tuffs and aerial breccias. These rocks not 
only form the cones and slopes of the volcanic mountains, but extend in 
certain regions over the plains and valleys, and mingle with the sands and 
coral limestones which make up the sedimentary rocks of the islands. 

2. Those igneous rocks which have solidified at some depth beneath the 
surface, and in places have been considerably metamorphosed by shearing 
and recrystallization, and in other ways changed ; and which have become 
exposed through erosion. These rocks appear to be older than the decent 
volcanic rocks, but may not be older than the earliest tertiary formations. 
They may be exposed intrusions of the volcanic lavas that form the older 
portions of the volcanic mountains mentioned in the first paragraph. 

3. The sedimentary deposits that have been derived from the igneous 
masses by disintegration in some instances, and by degradation of tuff 
deposits, or by direct sedimentation of such volcanic material at the time 
of its eruption, together with such other sediments as may have been 
derived from coral reefs, and other rocks. 

1. BOCKS OF THE VOLCANIC MOUNTAINS. 

The great bulk of all the volcanic rocks of the islands is andesite, for 
the most part pyroxene-andesite, in which hypersthene and augite are 
both present, the orthorhombic pyroxene often in great abundance. A 
considerable proportion of the andesites carry hornblende besides pyrox- 
ene, and are hornhlende-pyroxene-andesites. A smaller number are horn- 
blende-andesites without pyroxene, and fewer contain biotite in addi- 

^This work was performed in the laboratory of the division of mines, Bureau 
of Science, Manila. 

155 



156 IDDINGS. 

tion to the ferromagnesian minerals already named. Some pyroxene- 
andesites carry small amounts of olivine, and form transitions between 
andesite and basalt. 

Basalts with a variable amount of olivine are abundant, and constitute 
some of the more prominent and the more active volcanoes. On the other 
hand dacites, and possibly rhyolite, are rare, so far as known, and occur 
in relatively small bodies. 

Certain basaltic rocks characterized by rather alkalic feldspar, and in 
one case by altered leucite, are found in several localities. 

Pyroxene-andesite. — In most instances these rocks are dark colored, 
dense or porous, less often vesicular : porphyritic with many small pheno- 
crysts, that is, they are mediophyric. The relative amounts of pheno- 
crysts and groundmass vary somewhat in different cases, but the great 
majority have nearly equal amounts of phenocrysts and groundmass 
(sempatic), or have rather more groundmass (dopatic). Several of the 
freshest varieties collected may be described as follows : 

From Cochinos Point and Sisiman, Bataan Province, there is a dark-colored, 
sempatic, mediophyric rock; that is, one having many small phenocrysts, about as 
much in bulk as the groundmass which contains them. The phenocrysts are 
mostly labradorite, approximately AbaAn^, with pronounced zonal structure, the 
narrow outermost zone being distinctly alkalic. The shapes are those of rectan- 
gular prismoid to equant crystals. In size they are seriate; that is, of different 
sizes, from those of several millimeters to less than 1 millimeter. They are well 
twinned in Carlsbad, albite and pericline manner. They contain many micro- 
scopic inclusions, usually in the central portion of each crystal. There are fewer 
phenocrysts of hypersthene and augite, the former faintly pleochroic in thin 
section. Augite is twinned in some cases, and occasionally surrounds hypersthene. 
The pyroxene phenocrysts are euhedral, with the first and second pinacoids strongly 
developed. In size they are generally smaller than the largest feldspars. There 
is considerable magnetite in small crystals. Those inclosed in pyroxene are 
smaller than others not so inclosed. Some are inclosed in the margin of the 
feldspar. The groundmass consists of microlites crowded together; rectangular 
equant, also prismoid plagioclase feldspar, prismoid pyroxene, and equant mag- 
netite; probably with a cementing matrix of colorless glass. 

Almost identical rocks occur at Corregidor Island and Mount Arayat, 
Pampanga Province. In one rock from Corregidor the magnetite has 
crystallized in curved, skeleton-like form within a cluster of hypersthene 
and augite crystals, as when quartz forms graphic intergrowth with 
orthoclase; indicating synchronous crystallization of the magnetite and 
pyroxene. 

Pyroxene-andesites having the same composition and habit as that 
described from Cochinos Point occur in Albay Province, Luzon, on Cebu, 
and elsewhere in the Islands. 

A pyroxene-andesite of the same general type as those just described, 
that is, sempatic and mediophyric, with the same kinds of phenocrysts, but 
with a distinctly glassy groundmass, occurs at Siniloan, Laguna Province. 
The groundmass consists of brown globulitic glass, which is slightly more 



PETROGRAPHY OF IGNEOUS ROCKS OF PHILIPPINES. 157 

abundant than the euhedral, rather large microlites of rectangular, equant 
and prismoid plagioclase, prismoid pyroxene, and equant magnetite. 

Pyroxene-andesites with holoerystalline groundmass have been found 
in various localities: Benguet, Ilocos Norte, Masbate, Cebu, and else- 
where. In some the microlites are crowded together, without matrix of 
glass, and without any characters that distinguish the groundmass texture 
notably from many in which there is a slight amount of glass base, except 
that the form of the microlites is less distinct. In those with slightly 
coarser crystallization the fabric is characterized by anhedral, equant, 
micropoikilitic quartz, in indistinctly outlined anhedrons, crowded with 
minute prismoids of plagioclase and pyroxene, and anhedrons of mag- 
netite. 

A still coarser variety occurs at Mount Batalao, in southwestern Luzon. 

It is dosemjc, that is, the phenocrysts are more abundant than the groundmass; 
it is mediophyric, with phenocrysts between 1 and 5 millimeters in diameter. 
There is much labradorite in euhedral crystals, rectangular and tabular, with 
pronounced zonal structure, in recurrent zones of more or less alkalic com- 
position; well twinned, and containing many inclusions. There are fewer and 
smaller phenocrysts of pyroxene, both hypersthene and augite; and still less 
magnetite. The groundmass is holoerystalline, and consists of small euhedral 
plagioclase, less pyroxene and magnetite, with intersertal anhedrons of poikilitic 
quartz, crowded with minute inclusions, (chadacrysts) of plagioclase and alkali- 
feldspar, probably orthoclase. The intersertal quartzes are chadoikic, that is, 
there is about as much matrix quartz (oikocryst) as inclusions (chadacrysts). 
In a somewhat similar,, dosemic, mediophyric, pyroxene-andesite from 1 kilometer 
southeast of Aroroy, Masbate, the phenocrysts of pyroxene are larger than those 
of feldspar, and the larger augites inclose some plagioclase (labradorite), some 
hypersthene, and magnetite. The hypersthene is altered. The groundmass is 
itself porphyritic with small euhedral plagioclases in a matrix of consertal, 
anhedral crystals of more alkalic plagioclase, pyroxene, magnetite, and some 
quartz. This matrix is intersertal between the euhedral crystals. A similar 
fabric characterized a slightly altered pyroxene-andesite from Benguet, which is 
persemic; that is, has more phenocrysts than groundmass. 

An olivine-bearing pyroxene-andesite occurs at Antipole, Rizal 
Province. 

It is dopatic, with preponderance of groundmass; is mediophyric, and has a 
hyatal to seriate, porphyritic fabric. There are many phenocrysts of labradorite, 
with abundant inclusions; few of pyroxene and some olivine. The groundmass 
consists of colorless glass crowded with brown globulites, and many microlites of 
prismoid plagioclase, equant to prismoid pyroxene, equant magnetite, and many 
minute, long, needle-like crystals of pyroxene, and lines of globulites, which are 
undoubtedly pyroxene also. Some are in attached strings, some are disconnected. 
All the prismoid crystals are diverse in arrangement. This variety of rock is 
intermediate between andesite and basalt. 

Quite different types of texture from those just described occur in other 
varieties of pyroxene-andesite. 

One from Union Province is perpatic, with almost no phenocrysts, and these are 
of plagioclase. The groundmass consists of tabloid plagioclase, sections being 



158 IDDINGS. 

prismoid and equant, mostly indistinct bounded anhedrons, seriate in size, with 
some subparallel arrangement, or flow- structure. The feldspar is not notably 
birefringent except in prismoid sections, and is probably more alkalic than labra- 
dorite, possibly andesine. There are minute prismoids, or needles, of pyroxene, 
seriate in size, in places diverse in arrangement; also in clusters of subparallel 
prisms, crowded together closely by the side of the feldspar prisms or plates; 
also in slightly radiate bundles. Magnetite occurs in prominent and numerous 
euhedral and subhedral crystals. 

An exceptional rock related in fabric to that just described is found 
in gravel at Montalban, Rizal Province. 

It is a dense, greenish-gray, nonporphyritic rock, showing small megascopic 
crystals of feldspar and fewer of pyroxene. In thin section it consists of dendritic 
radiating clusters of nearly colorless monoclinic pyroxene, augite, and prismoids 
of feldspar. Also prismoid plagioclase in diverse arrangement and in many cases 
curved. The feldspar is probably oligoclase, as indicated by its optical properties. 
There is much green spherulitic serpentine or chlorite. There are small crystals 
of magnetite, and clusters, of brown anhedrons, probably epidote. There is some 
secondary quartz. 

Hornblende-pyroxene-andesites are very similar to pyroxene-andesites 
in composition except for the presence of variable amounts of hornblende, 
and generally smaller amounts of hypersthene. They resemble them also 
in texture, having similar habits and like variations in the fabric of the 
groundmass. There are varieties with little hornblende which grade by 
increasing amounts of this mineral and decreasing amounts of pyroxene 
into hornblende-andesites. A few rocks of this group contain small 
amounts of olivine. They are not so abundant in the Islands as pyroxene- 
andesite, judging from the collections already made, and occur intimately 
associated with them in various localities. 

Hornblende-pyroxene-andesite from Cabcaban, Bataan Province is dosemic 
seriate, and mediophyric; that is, it is richer in phenocrysts than in groundmass, 
and they vary in size from something less than 5 millimeters to microscopic 
crystals that may be considered as constituents of the groundmass. The phenocrysts 
are mostly labradorite, euhedral to subhedral. There are fewer of augite and 
hypersthene, and still less of hornblende, which is in relatively large crystals, 
greenish-brown, with dark borders. There are small phenocrysts of magnetite, and 
long thin prisms of apatite. The groundmass consists of microlites of plagioclase, 
pyroxene and magnetite, probably with a small amount of glass as matrix. There 
are some secondary minerals in the rock. 

Another variety from Dinalupihan, Bataan Province, is similar to that just 
described, but the phenocrysts of plagioclase carry fine glass inclusions; some of 
them brown, and in some crystals in zones. Brown hornblende without any dark 
border is abundant, and occurs also in small crystals in the groundmass. 

A hornblende-pyroxene-andesite from Subic contains euhedral crystals of 
labradorite (AbaAny), red-brown hornblende, with strong absorption and a narrow 
dark border, and an equal amount of pyroxene, mostly augite, besides magnetite. 
The groundmass consists of euhedral microlites of prismoid plagioclase and 
pyroxene, with a little hornblende and magnetite, with, probably, a small amount of 
colorless glass matrix. 



PETROGRAPHY OF IGNEOUS ROCKS OF PHILIPPINES. 159 

A somewhat similar rock occurs at Nagcarlang in Laguna Province; it is, 
however, dopatic, that is, it is richer in groundmass than in phenocryats. The 
euhedral phenocrysts of labradorite (Ab8An4) have some zonally arranged inclu- 
sions; there is much euhedral to subhedral, strong brown hornblende, without 
dark border ; in one case there is a shell of small pyroxene crystals. The pyroxene 
is augite in euhedral and subhedral crystals. The groundmass is brown glass 
crowded with euhedral microlites of rectangular, equant, and prismoid plagioclase, 
prismoids of pyroxene, a small amount of hornblende, and magnetite. 

Hornblende-pyroxene-andesite from Mount Batalao, Batangas Province, has 
red-brown hornblende with narrow dark borders, which occur also surrounding the 
pyroxene. The groundmass is holocrystalline and microgranular ; composed of 
indistinctly outlined, equant anhedrons, mostly feldspar. 

On Masbate, 1 kilometer southwest of summit of Mount Bagadilla, there is 
a variety with holocrystalline groundmass having the same fabric as the pyroxene- 
andesite already described; it is porphyritic, with small anhedral plagioclase in 
a matrix of consertal anhedrons of more alkalic plagioclase, pyroxene, magnetite, 
and some quartz. This matrix is intersertal between the small euhedral crystals. 

An olivine-bearing variety occurs on Mindanao. The hornblende is brown, the 
pyroxene is augite; there is a little colorless olivine, and possibly a little biotite. 
The groundmass is pumiceous and glassy, crowded with microlites of feldspar, 
pyroxene and magnetite. 

Horiiblende-andesite, with little or no pyroxene, is of widespread oc- 
currence. Its habit varies considerably, some varieties are magnophyric 
with abundant large phenocrysts of plagioclase about 10 millimeters in 
diameter, and smaller ones of hornblende. These are the rocks that were 
formerly called "trachytes.'' 

At Olongapo, Zambales Province, there is hornblende-andesite which is sempatic 
and seriate magnophyric. There are abundant phenocrysts of plagioclase and 
hornblende; few of pyroxene, and a few of quartz. The plagioclase is pronouncedly 
zonal with recurrent zones. The center is highly calcic, at least AbiAn4, bytownite. 
Successive zones are distinctly marked, with alternation in composition as shown 
by extinction angles; becoming generally more alkalic toward the margin, but the 
outermost narrow zone is strongly calcic like the central part of the crystal. 
There are many inclusions, usually clustered in zones. The hornblende is greenish- 
brown, contains many inclusions, often zonal, also central in some cases, with 
the outer portion of clear hornblende, and a narrow dark border. Paramorphs of 
hornblende are common. Rarely augite is inclosed, or partly surrounded by 
hornblende. The quartz phenocrysts are roimded. There are rounded colorless 
crystals of apatite. The groundmass consists of euhedral microlites of plagioclase, 
equant and rectangular, with small amount of matrix of alkalic feldspar, or 
glass base. 

The light-colored rock at Sisiman which is used in the Manila Break- 
water, is hornblende-andesite. 

It is sempatic, seriate and mediophyric. The most abundant phenocrysts are 
labradorite (AbjAna), euhedral and subhedral. They possess a narrow outer zone 
of distinctly more alkalic feldspar, which, however, has noticeably higher refraction 
than the anhedral feldspar of the surrounding groundmass. The hornblende is 
greenish -brown, but is mostly paramorphosed into aggiegates of magnetite and 



160 IDDINGS. 

pyroxene. There are few phenocrysts of pyroxene, and relatively large ones of 
magnetite. The groundmass is holocrystalline, composed of consertal anhedrons 
of feldspar, in part probably orthoclase, with some quartz. The rock is somewhat 
altered in parts, and contains calcite. 

The hornblende-andesite from Sta. Inez iron mines, Rizal Province, is typical 
of numerous others in these Islands. It is dopatic, seriate, mediophyric, with 
abundant euhedral phenocrysts of labradorite (AbgAn.O and many small pheno- 
crysts of greenish-brown hornblende with dark border. There are small magnetites, 
and a few colorless apatites, but no phenocrysts of pyroxene. The groundmass is 
holocrystalline, and consists of equant anhedrons of alkalic feldspar and quartz, 
with euhedral rectangular plagioclase, some minute anhedrons of magnetite, and 
brownish granules that. are probably pyroxene. Megascopically it appears to be 
distinctly crystalline, and represents a transitional variety between holocrystalline 
andesite and diorite-porphyry. Similar rock occurs in the Batanes Islands; in 
this the groundmass contains euhedral and subhedral quartzes that are micro- 
poikilitic. 

Homhlende-hiotite-andesite is found among the river gravel at Montalban. It 
is sempatic, seriate, and mediophyric. The plagioclase phenocrysts are euhedral, 
zonally developed, and approximately AbjAui. The biotite is brown; the horn- 
blende in this case is altered. There is magnetite and colorless apatite. The 
groundmass is microgranular, and consists of rectangular euhedral plagioclase 
with consertal anhedral feldspar, and micropoikilitic quartz. 

Dacites are not as common as the andesites and the few specimens 
collected are in part noticeably porphyritic^ in part megascopically non- 
porphyritic. 

One from Benguet is sempatic and hyatal, mediophyric; that is, there is a 
sharp contrast between the size of the smaller phenocrysts and the grain of the 
groundmass. There are many rather large phenocrysts of plagioclase (AbiAni) 
with slightly rounded exhedral forms, distinctly zonal; fewer and smaller 
phenocrysts of green hornblende, brown biotite, and still fewer of subhedral, 
rounded quartz; and some magnetite. The groundmass is microcryptocrystalline. 

A similar rock from Mindanao has the same habit ; but more phenocrysts, which 
are somewhat smaller, and more of them are ferromagnesian minerals. The 
somewhat rounded subhedral quartzes in some cases partly inclose euhedral 
brownish-green hornblende at the extremities of "bays" or pockets of groundmass, 
showing that these pockets and the rounded form of the quartz crystals were 
produced at the time of the crystallization of the quartz, and did not result from 
subsequent melting or solution of euhedral crystals. The groundmass is holocrys- 
talline, with consertal, equigranular fabric; composed of anhedrons of alkalic 
feldspar, in part probably orthoclase, and of quartz. This rock is a more quartzo^e 
hornblende-andesite. 

A similar type of dacite occurs on Masbate. The phenocrysts are similar, but 
the groundmass is slightly coarser, with much euhedral and subhedral crystals 
of quartz in a matrix of equant anhedrons of alkalic feldspar. These are also 
microcrystalline chlorite and epidote. 

On Corregidor Island massive dacite occurs which is white and dopatic, and 

. seriate, mediophyric. The phenocrysts are glassy plagioclase, quartz and biotite, 

megascopically black and euhedral. The groundmass is holocrystalline and seriate 

microporphyritic, consisting of small phenocrysts of prismoid plagioclase, probably 



PETROGUAPHY OF IGNEOUS ROCKS OF PHILIPPINES. 161 

andesine-oligoclase, and abundant, euhedral, bipyramidal quartzes,' in a matrix 
of anhedral quartz and intersertal alkalic feldspar. 

A similar dacite occurs east of the limestone, in the waterworks gorge at Mon- 
talban, Eizal Province. It is perpatic and mediophyric with few white phenocrysts 
in the specimen. The groundmass is coarser, and more granular, the euhedral 
and anhedral microscopic quartzes are somewhat larger than the minute prismoids 
of feldspar, and their outline is less sharply defined than those in the dacite from 
Corregidor, there is much minute subhedral magnetite, and globules of pyroxene, 
besides some chlorite. 

Megascopically nonporphyritic dacite, or possibly rhyolite, occurs in 
several localities. Its composition can not be determined by the micro- 
scope, because of the scarcity of the recognizable crystals. It is not 
known whether the feldspathic constituents are distinctly alkalic, or calci- 
alkalic. 

At Montalban the rock is microscopically porphyritic, seriate and dopatic, with 
microscopic phenocrysts of quartz, subhedral and also embayed, with inclusions 
of prismatic feldspar ; besides phenocrysts of alkalic plagioclase. The groundmass 
is microcryptocrystalline, with chlorite and iron oxide. Another variety at 
Montalban is green, and has the same kinds of microscopic phenocrysts. The 
groundmass is altered, brecciated, pumiceous glass which is now microcryptocrys- 
talline or microspherulitic. There is much chlorite in the pores of the pumice. 

On Cebu somewhat similar rock occurs. It has the fabric of collapsed pumice; 
that is, it consists of devitrified, welded, angular fragments of glass, now micro- 
cryptocrystalline. In it are small fragments of quartz, altered feldspar, and a 
few shreds of brown mica. 

One variety does not appear brecciated, but consists of a brownish-gray ground- 
mass, microgranular with indistinct spherulitic radial fabric. The granules appear 
to be minute anhedrons of quartz and alkalic feldspar with some magnetite. 
The spherulitic rays appear to be prismoids of ill-defined alkalic feldspar. These 
are small phenocrysts of sanidine. 

Bamlts. — There are transitions between olivine-bearing pyroxene-ande- 
sites and basalts rich in olivine, so there are basalts with the textures 
found in andesite, and others with texture not developed in andesites. 
ISTo line can be drawn between these two groups of rocks; and petro- 
graphers differ as to the classification of rocks intermediate between 
basalts and . andesites. It happens that the lavas of Mayon and Taal 
Volcanoes belong in part to these intermediate varieties, which may be 
called olivine-bearing pyroxene-andesites or andesitic basalts, while other 
varieties of lava from these volcanoes are normal basalts, with abundant 
olivine. 

Basalts with andesitic habit occur in Batanes. In two cases the rocks have 
the composition of hypersthene-augite-andesite with small phenocrysts of colorless 
olivine partly altered to iddingsite; the groundmass being holocrystalline, with 

^ These microscopic crystals of quartz were considered to be feldspar by Geo. F. 
Becker in his description of this rock. JJ. 8. Oeol. Surv. 21st Ann. Rept. (1901) 
Part 3, 516. 



162 IDDINGS. 

anhedral feldspars. These are probably best called oli vine-bearing pyroxene- 
andesites. 

Varieties with andesitic habit and much olivine occur at Mount Mariveles in 
Bataan Province and on Mindanao. In both of these rocks the olivine is present 
as small colorless crystals, in part altered to red iddingsite. In the Batanes 
Islands and Binangonan, Rizal Province, similar rocks have a groundmass with 
bro-svn globulitic glass base. In the first case it is filled with microlites of feldspar, 
pyroxene and magnetite. 

In the rock from Binangonan the groundmass is intersertal between the 
phenocrysts, and consists of the same kinds of minerals in smaller crystals with 
intersertal brown globulitic glass as a matrix. Somewhat similar basalts occur 
at Mount Arayat, Pampanga Province. 

In Lepanto there is a basalt which is dopatic and hyatal magnophyric, having 
a few large phenocrysts of olivine surrounded by small crystals of hypersthene. 
The groundmass is itself porphyritic, dopatic, seriate; that is, porphyritic with 
variously sized, small phenocrysts, which are plagioclase and some olivine. The 
groimdmass for these phenocrysts is holocrystalline, and consists of prismoid 
feldspars, in part with subparallel arrangement (flow structure), also smaller 
prismoid hypersthene and augite, rounded in form, and partly altered, together 
with much magnetite. The small ferromagnesian minerals occur between the 
crystals of plagioclase. There are also still smaller rounded pyroxenes scattered 
as inclusions through the feldspar. 

Basalt from Pagsanjan, Laguna Province, is dopatic, mediophyric, with few 
rather large clusters of labradorite (AKAna) containing many microscopic inclu- 
sions. There are smaller phenocrysts of red altered olivine. The groundmass 
consists of prismoid plagioclase, with smaller equant anhedrons of augite, 
subhedrons of altered olivine and others of magnetite, for the most part intersertal 
between the plagioclase, besides intersertal globulitic glass, the globulites appearing 
to be minute equant anhedrons of augite, together with minute prismoids of the 
same mineral. 

Somewhat similar basalts occur at San Pablo, Laguna Province, and on Min- 
danao. The first is minophyric, with small phenocrysts of augite, fewer of plagio- 
clase, and euhedral olivine with partial borders of minute magnetite crystals. 
The second is perpatic and minophyric. 

Basalt from the floor of the crater of Taal volcano is dopatic, mediophyric, 
with phenocrysts of subhedral, green augite, having inclusions in zones in some 
crystals; subhedral equant to tabular labradorite (AbjAna to AbiAnj), zonally 
developed; and fewer colorless to yellow olivines, altered on the surface of the 
crystals. The augite and labrodorite are anhedral toward each other when in 
clusters. The olivine in some instances is partly inclosed in augite, with anhedral 
forms. The groundmass consists of euhedral prismoid plagioclase, with central 
euhedral prismoid inclusion that has much lower refraction and is isotropic, 
apparently glass. The plagioclase prismoids have diverse arrangement. There 
is also much equant anhedral augite, less magnetite, and probably intersertal 
colorless glass, but the microlites are crowded close together, and the rock may be 
holocrystalline. Another variety of basalt from Taal Volcano is dopatic, medio- 
to minophyric, and. hyatal. The phenocrysts of plagioclase are euhedral and 
anhedral fragments, with many minute inclusions, and pockets and cracks 
filled with groundmass. In one crystal a narrow crack is filled with brown glass 
containing no microlites, whereas the glass of the surrounding groundmass is 
crowded with them, illustrating diflferential flow in a partly crystallized magma, 
and the probable origin of differential contemporaneous veins. There are few 



PETROGRAPHY OF IGNEOUS ROOKS OF PHILIPPINES. 163 

phenocrysts of olivine and none of augite in the thin section studied. The ground- 
mass consists of equant anhedral augite, much magnetite, less plagioclase, with 
very little brown glass, scarcely recognizable as such. 

Another rock from this region is perpatic, with few small phenocrysts of 
plagioclase, augite and olivine, in a groundmass like that in the second rock 
described from Taal Volcano, but finer grained. 

. Somewhat similar basalts with more olivine and nearly the same type of ground- 
mass occur on Mount Mariveles and elsewhere in Bataan Province. 

Basalt from Mayon Volcano, Albay Province, is dopatic and mediophyric, and 
highly vesicular or porous. The euhedral, rectangular, prismoid phenocrysts of 
laibradorite contain many inclusions of brown glass; the phenocrysts of green 
augite are subhedral; those of colorless olivine are subhedral to euhedral. The 
groundmass is a dark brown globulitic glass with microlites of thin needle-like 
prismoids of plagioclase, and anhedrons of augite and magnetite. 

Another rock from this locality, is dopatic, minophyric, with many phenocrysts 
of augite and colorless olivine^ and few of plagioclase. The groundmass is like 
that of the rock from Mayon Volcano, just described. 

A basalt from the quarry near San Pablo, Laguna, is quite different 
in texture from most of the others yet collected. 

It is saccharoidal, perpatic and minophyric, with small phenocrysts of olivine. 
The groundmass is holocrystalline, with microlitic cavities. It is chiefly prismoid, 
or bladed plagioclase in subparallel arrangement (flow structure), much anhedral 
olivine and augite, in small crystals. The olivine and plagioclase are in part 
anhedral with respect to each other, the olivine having begun to crystallize first. 
Magnetite occurs as minute anhedrons within the olivine, but mostly as intersertal 
and poikilitic anhedrons, partly surrounding, with rounded surfaces, plagioclase 
and augite, as the last component, or the chief eutectic component to end the 
crystallization; the intersertal orthoclase being the last feldspathic constituent 
to crystallize, probably synchronous with the magnetite. 

Basalts from Mindanao differ somewhat from those already described 
from Luzon in being richer in ferromagnesian minerals, and in having 
plagioclase slightly less calcic. 

One basalt from Mindanao is dopatic, mediophyric, and seriate; with many 
subhedral phenocrysts of olivine, but slightly altered^ and with inclusions of 
magnetite. The groundmass consists of much anhedral augite, some anhedral 
olivine, less magnetite, and prismoid plagioclase about equal to the ferromagnesian 
minerals in amount. The composition of the plagioclase is not readily determin- 
able, but it is as calcic as andesine.. There is a small amount of colorless matrix 
with lower ' refraction, which may be glass. 

A similar basalt from Mindanao occurs with smaller phenocrysts (minophyric) 
of olivine and augite. The groundmass is similar to that of the basalt just 
described, but is partly altered to a zeolite, or opal. 

Another basalt from Mindanao is porous, and like the last in habit. The few 
small phenocrysts are olivine, labradorite (AbjAn^), and pyroxene, in part hypers- 
thene. The groimdmass is chiefly pyroxene, with anhedral magnetite, and rod-like 
crystals of what may be ilmenite. - 

Still another basalt from Mindanao is perpatic, mediophyric, with few pheno- 
crysts of labradorite. The groundmass consists of euhedral prismoids of labra- 
dorite, equant anhedrons of augite, and equant subhedrons of olivine, in an 



164 IDDINGS. 

intersertal matrix of globulitic glass. The glass and crystals are in nearly equal 
proportions, and the average diameter of the crystals is about 1 millimeter. 

Somewhat similar basalt with much globulitic glass in the groundmass occurs 
on Palawan Island. 

On Masbate some of the volcanic rocks differ notably from most of 
those found elsewhere in the islands, in that they appear to contain higher 
amounts of potash, which shows itself in orthoclase, mica, and what was 
probably originally leucite. It is possible that related rocks will be found 
in other parts of the Archipelago when it is more thoroughly explored. 
These rocks have undergone considerable alteration and their exact com- 
position is not readily determinable from the sections studied. 

The rock from about 2 kilometers southwest of Aroroy is dark-gray, sempatic, 
mediophyric, and seriate; the diameters of the phenocrysts varying from 5 milli- 
meters to less than 1 millimeter. They are chiefly euhedral twinned augite, 
greenish with faint pleochroism, from green to yellow; clusters of equant colorless 
crystals, and some isolated euhedrons of a mineral which appears to be leucite, 
possibly altered to analcite, as it does not exhibit birefrigence. There are some 
small phenocrysts of plagioclase, partly altered and probably alkalic, or calci- 
alkalic. The augite phenocrysts are zonally developed, with "hour-glass" structure 
in some cases. ' They contain glass inclusions and small crystals of magnetite. 
Surrounding the phenocrysts, especially the augite, there are dark clusters of 
minute crystals that extend to various distances into the groundmass. They are 
needles and minute prismoids of augite in subparallel clusters. In places they 
seem to be granulated, or coarsely globulitic. With these prismoids are mingled 
opaque needles, or blades of what appears to be limonite; but from their resem- 
blance to similarly shaped and arranged blades of brown mica, present in closely 
related rocks of this locality, they may be paramorphic mica. They are in sets, 
or groups, of parallel prismoids, often at different angles in the matrix, or in 
fern-like arrangement; a set of minute, parallel blades being crossed by a 
larger one like a stem. 

The cluster of colorless minerals have the form of leucite crystals, are euhedral 
to subhedral; and in some cases rounded. Some carry minute inclusions, centrally 
located, less often zonally. In several finer-grain'ed varieties of the rock there 
are minute, imperfect, skeleton forms, characteristics of leucite. The groundmass 
consists of indistinctly outlined, clouded feldspar, possibly orthoclase in part, 
besides products of alteration. In the finer-grained variety the prismoid and 
needle-like alkalic plagioclase is more distinct. There are prismoids of augite, 
and groups of parallel needles or blades of pale brownish, pleochroic mica. 

Another rock from about 1 kilometer southwest of Mount Bagadilla is like the 
one just described in the habit of the groundmass, is but slightly porphyritic, and 
contains less ferromagnesian minerals. There are groups of minute blades of 
mica in parallel arrangement, and a few imperfect skeleton forms of leucite. The 
rock is partly altered. 

The rock from the south side of the Guinobatan River, south of Mount Baga- 
dilla, is an altered rOck like the one just described. A coarser-grained variety 
from the locality last mentioned is slightly porphyritic, with shreds, or blades, 
of light brown mica in diverse arrangement, and with a matrix of clouded 
orthoclase. It forms a transition between the leucitic rocks just described and 
the syenite. It is also somewhat decomposed. 



PETROGRAPHY pP IGNEOUS ROCKS OP PHILIPPINES. 165 

2. COARSER-GRAINED IGNEOUS ROCKS. 

Syenite. — ^A medium to fine grained rock occurs about 1 mile east of the summit 
of Mount Aroroy on Masbate. It is non-porphyritic and consists of rectangular, 
prismoid to equant, anhedral feldspars, that are to some extent twinned in the 
Carlsbad manner, are cloudy orthoclase, probably sodic. Their arrangement is 
diverse. There is a subordinate amount of ferromagnesian minerals in smaller 
crystals included in the feldspars, and also intersertal to them. They are chiefly 
bro^vn mica, partly chloritized, diverse in arrangement; some magnetite, and long 
thin needlelike crystals of apatite, with also diverse arrangement. There is some 
scattered chlorite, epidote and secondary quartz. 

Another variety of syenite on Masbate from the south side of the Guinobatan 
River, south of Mount Bagadilla, is like the one just described in texture, grain, 
and the character of the feldspars, but contains more ferromagnesian minerals; 
biotite in prismoids, or blades; pale green augite, partly replaced by calcite; a 
small amount of what appears to be serpentinized olivine; also magnetite and 
apatite. The rock approaches shonkinite in composition, but has rather more 
feldspar than ferromagnesian minerals. These syenites are undoubtedly related 
genetically to the leucitic lavas and are described in connection with them for 
this reason. Other syenitic rocks, chiefly albite, will be described after the 
commoner phanerocrystalline rocks. 

There are rocks in the collection that are normal gabbros with ineqni- 
granular consertal fabric, sometimes called "granitic/' and others with 
ophitic fabric, also called ''diabasitic.'' They grade through finer ^grained 
porphyritic varieties into holocrystalline pyroxene-andesites, and basalts. 
They also grade into diorites and quartz-diorites by variations in the 
mineral composition, just as the basalts grade into pyroxene-andesites, 
and these into hornblende-andesites and dacites. The extremes of the 
mineral variation, or differentiation in this region appear to be albitic 
granite and albitic syenite on the one hand, and peridotite and possibly 
pyroxenite on the other. The alteration of pyroxene into fibrous am- 
phibole, uralite, and also into compact hornblende in some instances, 
leads to the metamorphism of some of the gabbros into metadiorite, which 
in such cases is usually recognized by the character of the feldspar and 
the texture of the rock. 

There are not sufficient data at hand to determine the relative abun- 
dance of the different magmas and so indicate the composition of the 
average or "parent'^ magma. It appears that the coarser grained, in- 
trusive rocks have attained a higher degree of differentiation than the 
extrusive lavas, but this may not be the actual fact, and further study of 
the region may modify this conclusion. 

The best known active volcanoes are erupting lavas, but are not ex- 
tremely differentiated. The modern lavas are not all alike, and some of 
the older ones are highly differentiated from the probable parent magma, 
as, for example, the dacite on Corregidor and the leucite rock on Masbate. 

Gahhro. — The gabbros with consertal, inequigranular fabric vary from 



166 IDDINGS. 

those with nearly equigranular feldspars to others with intersertal fabric, 
due to the abundance of smaller crystals between the larger ones. There 
is thus a transition to seriate porphyritic fabric, and the texture of 
holocrystalline andesites. The similarity of the component minerals also 
indicates a genetic relation between these coarser grained rocks and the 
extrusive lavas in this region. Some of the freshest examples in the col- 
lection will be described. 

An blivine-gabbro occurs in river gravel at Montalban. It is medium to coarse 
grained, inequigranular, consertal; and consists of labradorite (AbaAua), pale 
greeh augite, colorless olivine, and very small amounts of primary hornblende, 
biotite and magnetite with some secondary minerals locally developed. The 
proportion of feldspar to ferro-magnesian minerals is about 3 to 1. 

The labradorite is anhedral, without zonal structure, is somewhat twinned 
according to the three common laws, albite, pericline and Carlsbad; and contains 
a small amount of minute rod-like inclusions. The augite is anhedral; has a 
distinct pinacoidal cleavage characteristic of diallage; is twinned, and contains 
rod-like inclusions. The olivine is also anhedral; in places it is intersertal with 
respect to the labradorite, showing its later crystallization. There is a narrow 
border of minute anhedrons, probably amphibole, between the olivine and labra- 
dorite. The proportion of augite to olivine is about 6 to 1. Primary green 
hornblende surrounds augite and magnetite in some places, and light brown biotite 
in very small amount is associated with the hornblende. The magnetite is 
associated with the augite, in juxtaposition and also intersertal between augite 
anhedrons. The intersertal position of much of the augite, olivine, magnetite 
and primary hornblende between the larger anhedrons of labradorite indicates 
the synchronous crystallization of the ferro-magnesian constituents together witli 
some labradorite, as the last act of the crystallization of the magma, after much 
feldspar had crystallized. In places this rock is altered, and secondary green 
hornblende has been formed in narrow veins traversing altered feldspar and other 
minerals. This shows the ordinary metamorphism to metadiorite. Olivine-gabbro 
similar to that just described in texture and composition occurs in Albay Province. 

A coarser-grained olivine-gabbro of this type occurs in Nueva Vizcaya. The 
labradorite is more abundant, and is about five times as much as the olivine and 
scarce augite. There is no magnetite, but some secondary amphibole. 

At Nagcarlang, at the base of Banahao Moimtain, Laguna Province, there is 
gabbro having the same type of texture as that just described. It is without 
olivine or magnetite, and consists of monoclinic and orthorhombic pyroxene with 
some pale green secondary hornblende, and a smaller amount of labradorite. 

At Subig, in Zambales Province, there is gabbro with mediimi grain, inequi- 
granular, seriate, consertal fabric, composed of approximately 5 parts of labra- 
dorite and 1 of pyroxene, without olivine or magnetite. There are some stout 
prismoids of labradorite surrounded by equant anhedrons of the same mineral. 
Some of the pyroxene is irregularly shaped, relatively large anhedrons partly 
surrounding labradorite, while other pyroxenes are intersertal to the feldspar, and 
some are small anhedrons. All of the pyroxene is uralitized. 

A medium-grained gabbro occurs in Benguet. Its fabric is inequigranular 
consertal, with rectangular subhedral prismoids of feldspar and pyroxene in 
diverse arrangement. The feldspar and ferro-magnesian minerals are in nearly 
equal proportions. The feldspar is labradorite (AbjAnj), without zonal structure. 



PETROGRAPHY OP IGNEOUS ROCKS OF PHILIPPINES. 167 

The pyroxene is in subhedral stout prismoids, and is pale green augite with com- 
pletely altered orthorhombic pyroxene in smaller amount. The augite is also 
partly uralitized. Magnetite occurs in rather large subhedrons. There is a very 
small amount of anhedral quartz, intersertal to the other minerals. 

On Grande Island, Subig Bay, there is norite, or gabbro rich in hypersthene. 
It is fine-grained, inequigranular, consertal, and some varieties are distinctly 
seriate with a slightly intersertal fabric. The feldspar and ferro-magnesian 
minerals are in nearly equal proportions. The feldspar in labradorite (AbjAng). 
The pyroxene is mostly hypersthene; some crystals exhibiting laminated inter- 
growth with monoclinic pyroxene. Green hornblende forms borders about some 
crystals of hypersthene. The magnetite is anhedral and partly surrounds labra- 
dorite and hypersthene with curved forms, as a synchronous and also a later 
crystallization. In the variety of the rock with intersertal fabric, the smaller, 
intersertal crystals are labradorite, pyroxene, hornblende and magnetite. The 
relative positions and shapes of the minerals indicate that while there was 
synchronous crystallization, labradorite was the first to begin, then pyroxene, 
and lastly magnetite. This rock has the same mineral composition as some of the 
pyroxene-andesites in the neighboring region, and the larger crystals are about 
the size of the phenocrysts in many of these lavas. 

A very fine-grained norite occurs on Palawan, having the composition approxi- 
mately of 50 per cent feldspar, 40 per cent pyroxene, 10 per cent magnetite. The 
fabric is nearly equigranular consertal. The labradorite is anhedral and somewhat 
larger than the crystals of pyroxene. They contain small inclusions of euhedral 
pyroxene and magnetite. The pyroxene is mostly hypersthene, in nearly equant 
subhedrons and rounded anhedrons. 

Gabbros with ophitic fabric which are sometimes called "diabase/^ 
dolerite, or phanerocrystalline basalt^ occur in numerous localities in the 
Archipelago, but most of those already collected are more or less decom- 
posed, or metamorphosed. They are characterized by prismoid, or tabular 
plagioclase, with diverse, less often subparallel, arrangement; and by 
poikilitic, or intersertal, pyroxene that acts as a matrix for the feldspar. 
They grade into varieties in which the intersertal matrix is formed of 
several crystals of more than one mineral, as in some basalt. 

The ophitic gabbro, or dolerite, from Malirong Falls, Leyte, is ophitic, and 
seriate, with about equal amounts of feldspar and ferro-magnesian minerals. The 
plagioclase is altered. The augite is in part poikilitic, and is colorless in thin 
section, with a purplish outer zone. There is considerable magnetite which is in 
part intersertal between the crystals of augite and feldspar. There are some 
chloritized, or serpentinized, crystals which may have been olivine. They are 
subhedral in form, and quite abundant. 

Other rocks of this kind occur in gravel in the Baliuag River, Bulacan Province ; 
some varieties being coarse-grained, others extremely fine-grained, they also occur 
in the river gravel at Montalban. A dolerite, from Angono, Rizal Province, has a 
fabric related to ophitic, and consists of rectangular prismoid plagioclase in 
diverse arrangement, with much less abundant augite, in euhedral and subhedral 
prismoids, and also in anhedrons intersertal to the feldspar crystals; the augite 
is colorless with a brownish outer zone. Magnetite occurs in relatively large 
skeleton-like anhedrons. There are some serpentine pseudomorphs of olivine, and 
considerable chlorite. 
93217 8 



168 IDDINGS. 

Metadiorite, which appears to be metamorphosed gabbro, is of frequent 
occurrence. 

That from Malitbog, Leyte, is medium-grained, inequigranular and consertal, 
and consists of labradorite, without zonal structure, and green hornblende, both 
compact and fibrous. 

Another metadiorite occurs at Subig, Luzon. It is medium to fine grained, 
inequigranular, consertal, and slightly intersertal. It consists of zonal labradorite, 
with a marginal zone of more alkalic plagioclase and some intersertal areas of 
graphic quartz in alkalic feldspar. The pyroxene is subhedral, and is completely 
utilized. There is some secondary compact amphibole, and considerable magnetite. 
The rock is an altered quartz-bearing gabbro. 

Diorites, which are plagioclase rocks characterized by notable amounts 
of primary hornblende, are related to hornblende-gabbro on the one hand, 
and to quartz-biotite-diorite, on the other. They may be confused with 
metadiorite, in which the hornblende is not pyrogenetic. 

Near Atimonan, Tayabas Province, there is a medium-grained, inequigranular, 
consertal diorite, which may possibly be hornblende-gabbro. The labradorite is 
like that common in gabbro; without zonal structure, and containing numerous 
rod-like inclusions. The hornblende is anhedral and green, with numerous rounded 
inclusions. There is a small amount of mica altered to chlorite. Magnetite is 
in part intersertal. 

Hornblende-pyroxene-diorites are found in the river gravel at Montalban, Rizal 
Province. They vary in texture from that of normal diorite to that of holocry- 
stalline andesite, indicating that these diorites are coarser-grained intrusive 
andesitic magmas like those erupted as lavas from the volcanoes. 

One variety is medium to fine grained, with inequigranular, consertal, seriate 
fabric, with intersertal portions. The plagioclase is labradorite considerably 
altered, somewhat zonally developed, in equant and tabloid anhedrons. The horn- 
blende is brownish, and anhedral. Uralitized pyroxene is present, and magnetite. 
There is much chlorite, epidote and some quartz. 

Another specimen, which is altered, contains much labradorite, colorless augite, 
hornblende and a small amount of brown biotite and magnetite, Still another 
variety from the same gravel is slightly coarser, with the same type of fabric, 
but with larger areas that are intersertal between feldspar crystals. These inter- 
sertal portions are in part poikilitic, or graphic, pyroxene and green hornblende 
anhedrons. The inclosed mineral is alkalic feldspar without striations, orthoclase 
or albite. In some instances the pyroxene and hornblende are intergrown with 
each other. The augite is colorless and abundant, and is not all poikilitic. The 
hornblende is pale green to colorless. 

A finer-grained variety from Sta. Inez, Rizal Province, is seriate porphyritic, 
with small amount of groundmass. The phenocrysts of plagioclase (AbiAui) are 
rectangular, prismoid or bladed euhedrons and subhedrons, with zonal structure; 
the outer zone being more alkalic. The hornblende and pyroxene phenocrysts 
are anhedral. The groundmass consists of rectangular prismoid and equant 
euhedral plagioclase, with some anhedral quartz, and a small amount of alkalic 
feldspar, probably orthoclase. There is brownish-green hornblende in the ground- 
mass which is poikilitic with small chadacrysts of feldspar. There is a small 
amount of altered mica in the rock. This rock is a fine-grained hornblende- 
pyroxene-diorite, or holocrystalline honnblende-pyroxene-andesite. 



PETROGRAPHY OF IGNP^OUS ROCKS OF PHILIPPINES. 169 

Diorites rich in hornblende occur in Cebu. One from the Island of Palawan 
corresponds in general texture to the fine-grained norite already describd from 
this island. Its fabric is consertal, anhedral, equigranular, or nearly so. It 
consists of nearly equal amounts of labradorite and brownish-green hornblende. 
Within the feldspar are small euhedral inclusions of hornblende and magnetite; 
and within the hornblende anhedrons there are small anhedral inclusions of 
feldspar and magnetite. 

Quartz-diorites occur in different parts of the Islands. 

At Antamok, Benguet Province, there is medium-grained quartz-diorite, with 
inequigranular consertal fabric. It consists of plagioclase and considerable 
brownish-green hornblende, anhedral with respect to each other, but euhedral 
toward quartz and orthoclase. There is some altered biotite. In places the 
orthoclase is inter sertal to poikilitic, with inclusions of plagioclase and hornblende. 

Another quartz-diorite from Camarines Province has the same texture as that 
just described, but there is less hornblende, and more quartz. The plagioclase 
is probably zonal andesine, with the outermost zone more alkalic. 

Other quartz-diorites have been found near Talisay and in the Loboo Mountains, 
in Batangas Province, on Masbate, on Lepanto, and elsewhere. 

Near Sara, Iloilo, there is quartz-diorite which is more or less porphyritic and 
forms a transition between non-porphyritic fine-grained quartz-diorite and holo- 
crystalline dacite. One variety is fine grained and slightly porphyritic; that is, 
perpatic, the inconspicuous phenocrysts, about 5 millimeters in diameter and less, 
being plagioclase. The groundmass, or chief portion of the rock, is inequigranular, 
and consists of euhedral, rectangular prismoid and equant plagioclase, zonally 
developed; with euhedral, long prismoids of brownish-green hornblende, and con- 
siderable magnetite; together with anhedral poikilitic quartz and orthoclase, which 
form about 15 per cent of the whole rock. There is a small amount of sphene, 
intersertal with respect to the plagioclase, a very uncommon mode of occurrence. 

Another variety of this rock, which appears megascopically to be medium- 
grained and non-porphyritic, is in fact dopatic mediophyric; that is, there is 
much groundmass between the medium-sized phenocrysts. It is hyatal. The 
phenocrysts are euhedral, rectangular prismoid plagioclase, subhedral and some- 
what rounded quartz; euhedral prismoid hornblende and relatively large subhedral 
magnetite. The groundmass is micrographic, quartz and clouded alkalic feldspar, 
probably orthoclase. There is some secondary quartz and zeolite filling what were 
possibly microlitic cavities. 

On Masbate there are porphyries intermediate between porphyritic 
quartz-diorite and holocrystalline dacite. 

One variety, from the Guinobatan district, Aroroy, is mediophyric sempatic and 
seriate, with wide range in the sizes of the phenocrysts. These of plagioclase are 
euhedral, with jagged outline due to the interference of the crystals in the 
groundmass. They are notably zonal. The ferro-magnesian phenocrysts have been 
completely chloritized, and were mica, and possibly hornblende. The groundmass 
consists of small plagioclase and a matrix of much quartz, both euhedral and 
anhedral, with some orthoclase intersertal between the quartzes. There is 
magnetite in small amount. 

Another porphyry from the west of Buyuan Bay, Masbate, is much like that 
just described. Similar rocks occur in Benguet. 



170 IDDINGS. 

Granite is rather an uncommon rock, so far as present observations go, 
and the few bodies of granite known are somewhat metamorphosed, or 
altered. 

That found near Paracale, Ambos Camarines, has a gneissoid texture, is medium 
to fine grained, and consists of white feldspar and quartz and darkgreen chloritized 
mica. It is inequigranular, with equant anhedral feldspar and quartz crystals, 
with smaller amounts of intersertal anhedrons of the same minerals, besides 
chloritized mica and epidote. The feldspars are orthclasfe, and less oligoclase or 
albite. There is considerable quartz. 

The granite at Mambulao has been sheared to a thinly laminated gneiss with 
^'augen-struktur*^ on a small scale. The orthoclase and albite lie as anhedral 
blocks in a matrix of smaller equant anhedrons of quartz and orthoclase, with 
shreds of muscovite, having pronounced fluxion structure. 

Peridotites occur at various localities on Luzon in association with 
gabbro, and probably some pyroxenite. They are mostly altered to ser- 
pentine. 

On Grande Island, Subig Bay, there is medium-grained periodite with consertal 
fabric. The component crystals are anhedral, and consist of pale-green augite, 
and faintly pleochroic hypersthene, colorless olivine, some green hornblende inter- 
sertal with respect to the other ferro-magnesian minerals, and some plagioclase, 
at least as calcic as labradorite. The two kinds of pyroxene are interlaminated in 
some crystals, and some hypersthene is poikilitic with plagioclase inclusions. 
There is no magnetite. 

Another peridotite quite similar to that from Grande Island, occurs at Malitbog, 
on Leyte. It contains much partly serpentinized olivine; also some dark brown 
intersertal mineral in small anhedrons, which appears to be garnet. 

Highly serpentinized peridotite occurs at Mambulao, Camarines Prov- 
ince, and also in Ilocos Norte, in Albay, and elsewhere. 

In Ilocos Norte there are chlorite rocks; some with epidote, others with 
rutile, actinolite, and muscovite. Their origin is not at present known. 

On Mindanao there is a laminated, sheared rock composed of fragments 
of hornblende and pyroxene, without feldspar, which appears to be sheared 
pyroxenite. 



THE PHILIPPINE 

Journal of Science 

A. Chemical and Geological Sciences 
AND THE Industries 



Vol. V MAY, 1910 No. 3 



THE STUDY OF MANILA COPAL. 



By Paul C. Freer. 
{From the Bureau of Science, Matiila, P. L) 



The chemical laboratory, assisted by the botanists -of the Bureau of 
Science, has for some years past been engaged in the study of minor 
forest products, as time and materials permitted. Some of this work 
has already appeared in several papers on resins and wood oils, chief 
among which have been those on Manila elemi} 

Probably the main article of export of this nature from the Philippine 
Islands is Manila copal, the Spanish term almaciga for many years 
having been locally applied erroneously both to it and the tree from which 
it comes. An expedition from this Bureau, as early as 1903, while 
ascending Mount Pulgar in Palawan, camped on the side of the moun- 
tain at an altitude of about 850 meters, in the midst of a magnificent 
forest composed almost entirely of a growth of Agathis alha (Lam.), 
from which tree the copal of commerce is derived. We observed fresh 
resin, sometimes in large masses, on the sides and in the notches of the 
trees wherever they had been wounded, but what was especially interest- 
ing, on digging along the roots of several especially large specimens, 
we observed not only fairly large masses of fossil resin contiguous to them, 
but small drops or masses of copal which had exuded from the smaller 
radicles extending deep into the ground- It appeared probable that 
all of these exudations were derived either from radicles which were 
dying or which had been injured in some way, perhaps by the bites of 
insects. Nevertheless, this observation led us to the conclusion that 
it is not by any means certain that injury must always precede the 

^Clover, A. M. Philippine Wood Oils. This Journal, 8eG. A (1906), 1, 191; 
The Terpene Oils of Manila Elemi. (1907), 2, 1. Bacon, Raymond F. Philippine 
Terpenes and Es^ntial Oils, I. This Journal, Sec. A (1908), 3, 49; Philippine 
Terpenes and Essential Oils, III. (1909), 4, 93; Philippine Terpenes and Essential 
Oils, IV (in press). Richmond, George F. Philippine Turpentine. Editorial. 
This Journal, Sec. A (1909), 4, 231. 

95193 . 171 



172 FREER. 

exudation of the resin, although Mr. Richmond, in subsequent careful 
investigations of the trees, concluded that such must be the case. 

The study of amorphous bodies, such as resins, which in by far. the 
greater part yield only amorphous decomposition products, is naturally 
attended with great difficulty, and consequently the publication of the 
results from this laboratory has been delayed for a long time. The 
work was begun by Dr. Penoyer L. Sherman, and continued by Mr. 
George F. Richmond and Mr. Benjamin T. Brooks, the botanical work 
being by Dr. F. W. Foxworthy. 

Manila copal is for the greater part composed of amorphous resin 
acids, with but a small proportion of a neutral, amorphous solid, whicli 
is left behind on exhaustive treatment with alkalies. It is this latter 
body which darkens rapidly on melting and which, in all probability, 
is the cause of the deepening of the color of the resin when it is heated 
for varnish making. The pure resin acids, on heating, do not darken 
very much. Maaila copal also contains a certain, but small percentage, 
of terpenes. It is interesting to note that, whereas Manila copal, per 
se, rapidly oxidizes on exposure to the air, the pure acids, as isolated by 
Mr. Richmond, could be redissolved, precipitated, and even distilled with- 
out apparent change, the oxidation phenomena appearing to take place in 
the original resin. 

Copal gives off carbon dioxide even at ordinary temperatures if it is 
exposed to the air, and more rapidly in the sunlight. Undoubtedly 
this phenomenon is due to the formation of peroxides, and whereas it 
may be of importance in the finely powdered substance, it can not be 
of such great moment in the alterations which take place in the solid 
pieces, as otherwise the latter would have suffered profound changes and 
.would, according to the length of their exposure to air, differ markedly 
in their composition and properties, a fact which does not appear from 
the work which has been carried on in this laboratory. It is well known 
that the resin hardens promptly when it exudes from the tree, but this 
effect has nothing to do with oxidation. The small percentage of carbon 
dioxide which passes off at lower temperatures when copal is heated also 
shows that no profound oxidation has taken place. 

The resin acids are therefore present in the exudate as it issues from 
the tree. The analyses which at present exist give no prOof that these 
acids are closely related to each other, neither has anything been done 
which can clear up their constitution. The one crystalline acid, CjoHigO., 
forms only a small proportion of the whole, and although conjectures as 
to its nature are obvious, much more work needs to be done upon it. 

The studies which follow show some advances on the present methods 
of producing oleoresinous varnishes by outlining the means by which 
they may be prepared at lower temperatures, and they show the way by 
which the local trade could be supplied with satisfactory varnishes mado 
entirely from Philippine materials. 



THE ALMACIGA TREE: AGATHIS ALBA (LAM.) 

3 

i 
By F. W. FoxwoBTHY. . 

{From the botanical section of the Biological Laboratory, Bureau of Science, 

Manila, P. /.) 



The almaciga tree, Agathis alba (Lam.), was first described at length 
by Eumphius^ under the name of Dammara alba which is merely the 
Latin form of the Malay common name ''Dammar putL" In 1786 
Lamarck ^ gave a review of Eumphins^s work, using the name Dammara 
alba. In 1803, Lambert^ figured and described as Pinus dammara 
specimens of the same tree from Amboina. In 1824 he adopted the 
name Dammara orientalis. In 1807, Salisbury* proposed the name 
Agathis in place of Dammara, calling his plant Agathis loranthifolia. 
Other names which have been given to it are Abies dammara Poiret,*^ 
Agathis dammara Eich.,^ and Dammara rumphU Presl.'' 

Agathis is one of the nomina conservanda of the Vienna Botanical 
Congress of 1905. The name of this species, therefore, is Agathis alba 
(Lam.). This combination seems to have been used, since E. C. Jeffrey, 
in an article on "The Wound Eeactions of Brachyphyllum/'^ uses the 
name Agcithis alba without comment. 

Warburg^ maintains that this species is an aggregate form and he 
proposes to separate the following: 

A. dammara (Lamb.) Eich., the South Moluccas. 

A. lahillardieriW^Th., Amboina (?) or Waigu (?). 

A. regia Warb., the Moluccas, Batjan and the small island of Pulo 
Dekat. 

A. macrostachys Warb., Java and Sumatra. 

* Herbarium Amboinense (1741), 2, 174, pi. 57. 
2 Encyclopedie (1786), 2, 259. 

•''A Description of the Genus Pinus (1803), 1, pi. 38. 
^ Trans. Linn. 8oc. (1807), 8, 311. 

* Encyclopedie. Supplement (1817), 5, 35. 

•Commentatio botanica de Coniferis et Cycadeis (1826), 83, t. 19. 
^ Epimeliae BotanicflB (1861), 236. 
^Ann. Bot. (1906), 20, 387. 
•Monsunia (1900), 1, 182-185. 

173 



174 FOXWORTHY. 

A, rhomhoidalis Warb., the Malay Peninsula. 

A, beccarii Warb.^ Borneo. 

A, horneensis Warb._, Borneo. 

A, philippinensis Warb., the Philippines. 

A, celehica (Koord.) Warb., Celebes. 

A, motleyi (Pari.) Warb., Borneo. (Since credited to Podocarpiis.) 

Differences in size arid shape of leaves, size of staminate and pistillate 
cones, and appearance jof the resin are made the basis of this division. 
After comparison of fresh and dried material of A. philippinensis, A. 
beccariij A. horneensiSj smd a study of. material of A. celehica kindly 
loaned me by Dr. Treub, I am convinced that these four species at any 
rate should not be separated from the original A, alha. The tree is a 
very large one and material taken from different parts of the same 
individual shows a wide range of variation. A single Philippine tree 
will yield material to fit the descriptions of A. dammara, A. philip- 
pinensis, A, celehica, A. heccarii, and A. hormeensis, as given by Warburg. 
' I have studied the Philippine and Bornean trees in the field and they 
are identical in habit. Doctor Beccari ^^ also considered A. heccarii and 
A. horneensis merely varieties of A. alia. 

Described differences in the appearance of the resin may be due to the 
different seasons of collection, or to the different age of the samples 
collected. While by no nieans prepared to say what is the status of 
the genus in other parts of Malaya, I can see no reason for supposing 
that there is more than one species in the Philippines, or that it is 
other than Agathis alha. 

This is closely related to the Kauri, Agathis rohusta (Moore) Warb., 
of Australia and New Zealand. Other species of the genus are found 
in 'New Caledonia and Fiji. 

The almaciga is a very large tree, growing to a height of from 50 to 
60 meters and, to a diameter, breast high, of more than 2 meters, with a 
clear length of 30 meters or more. The bark is rather smooth and of a 
grayish color. It is a centimeter or more in thickness and contains 
numerous longitudinal resin canals. The tree has an altitudinal range 
of from 150 to 2,000 meters above sea level, attaining its best develop- 
ment in the Philippines on well-drained slopes at 600 to 1,500 meters 
above sea level. It is known in the Philippines by the following 
names: Almaciga (Spanish) ; hiayo, hidyangao (Visayan) ; hunsog, litao 
(Igorot) ; dinwr (Bagobo) ; galagala (Tagalog) ; ladiangao (Bicol) ; 
macao (Moro) ; saleng, uli (Negrito). 

The distribution in the Philippines is as follows: Island of Luzon, Provinces 
of Cagayan, Lepanto, Benguet, Zambales, Bataan, Camarines, Sorsogon, Albay. 
It is abundant on the Islands of Mindoro, Negros, Palawan, Sibuyan, and Min- 
danao, on which last-named island it is kno^vn from Misamis, Davao, and Zam- 

'" Nelle foreste di Borneo, 163. 



ALMACIGA tree: AGATHIS ALBA (LAM.), 175 

boanga. It is probable that it is more widely distributed than is shown by our 
collections. It is found in the extreme north of Luzon and in the most southern 
islands of the group. 

This tree is the source of almaciga or Manila copal, which is a con- 
siderable article of export from the islands. The almaciga seems to 
be the only product of this tree which is used. It is employed locally 
for incense in religious ceremonies. The resin is found in the bark, as 
already indicated, and it oozes out wherever the bark is cut. Oc- 
casionally, lumps of hard resin are found in the forks of branches and 
sometimes the so-called fossil resin is encountered in masses in the ground 
at the base of the tree. The gathering of the resin is the principal 
occupation of some of the Tagbanuas of Palawan. 

The dipterocarp resins are not gathered to any considerable extent 
in the Philippines; consequently, there is not much likelihood that the 
resins of the Dipterocarpacece will be mixed locally with Manila copal. 

Doctor Beccari met with this tree on the upper slopes of Mount Poe, Sarawak. 
He says" that the resin collects at the foot of the tree and forms stone-like 
masses. There he heard it called ''Dammar daghin*' or "flesh-resin," and con- 
sidered it to be one of the best resins. I collected specimens from the tree in 
the same locality as Doctor Beccari and found the Land Dyaks terming it "Dama 
Undang." I did not see any of the deposits at the base of the tree, but I learned 
that the Land Dyaks make a business of collecting the resin from the tree. I 
found one tree which had a ladder on it made by driving pegs at intervals of about 
1 meter and tying saplings to these. This, I was told, is a common means of 
enabling the dammar hunter to get at the clear lump resin which is found at 
the forks or on the branches. 

The Sarasins " describe the same method of collecting Agathis resin in Celebes 
and say that the bark is also cut into to produce a flow of resin. 

^^ Log. cit. 

^^Reisen in Celebes (1905), 1, 182. 



MANILA COPAL. 



By George F. Richmond. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. /.) 



CONTENTS. 

Manila copal. 

The chemical examination of Manila copal. 

Oleoresinous varnishes. 



INTRODUCTION. 



This investigation was undertaken with a twofold object, first, of 
clearing up the considerable confusion which exists throughout the liter- 
ature and in commerge with respect to the origin, mode of production, 
commercial grading, and physical and chemical properties of this com- 
modity, and, second, with the hope that some light might be thrown 
upon the chemistry of the art of its most important use, namely, varnish 
making. "Manila copal,^^ so called because of its similarity to the true 
copals, as an ingredient of oleoresinous varnishes is a leading minor 
article of export the Philippine Islands, approximating 200,000 pesos 
($100,000 United States currency) in value annually. 

It appears in the Quarterly Summary of Commerce under the term 
^'^almaciga,^^ which is a Spanish name meaning "gum mastic,^^ a resin 
of entirely different character produced by Pistacia lentiscus Linn, in 
Spanish and other Mediterranean countries. Almaciga is the only 
resinous product exported from Manila and is identical with the Manila 
copal of commerce. 

GEOGRAPHICAL ORIGIN. 

According to Wiesner,^ Manila copal is a collective term used com- 
mercially for medium-hard resins from the Sunda, Philippine, and 
Molucca Islands, which show by their properties that they possess a 
common origin. Hence, Celebes, Singapore, Borneo, Philippine^ and 
East Indian gums are known to varnish makers, and. all of these in point 
of g^graphical origin may be included under the collective term of 
Manila copal. 

*Die Kohstoffe des Pflanzenreiches. Leipzig, 2d ed. (1900), 1, 284. 

177 



178 RICHMOND. 

BOTANICAL ORIGIN. 

Comparatively little is known regarding the source of the Manila copal 
of commerce . except that it is undoubtedly of vegetable origin and is 
collected in Malayan regions. So far as the Philippine resin "almaciga" 
is concerned, it is produced by a large coniferous tree, Agathis alha 
(Lam.), and Foxworthy^ has shown that an individual of this species 
of Agathis will furnish botanical material to fit the description of Agathis 
dammara Rich, of the Moluccas, Agathis celehica Warb. of the Celebes, 
Agathis heccarU and A. horneensis Warb. of Borneo, and Agathis philip- 
pinensis Warb. of the Philippines, thus combining five of the nine separate 
species of Agathis proposed by Warburg, and it seems reasonable to 
assume that a comparison of botanical material of this genu^ from other 
Malayan sources would further limit the botanical origin of Manila copal. 

It is thus seen how the prevailing custom of naming a natural product 
after the various localities where it is collected or after ports of shipment 
would tend to obscure the fact that the Manila copal of commerce, whether 
coming from Manila or elsewhere, .has a common botanical origin. 

AGATHIS ALBA (LAM.). 

Botanical description. — The Manila copal tree attains a height of 50 to 60 
meters and a diameter of 1 to 2 meters. The bark is thick and smooth, the 
wood is light colored and contains nmnerous resin cavities*. The leaves are oveate- 
lanceolate to elliptical in shape, 2 to 12 centimeters long and from 1 to 4 centi- 
meters broad, mostly opposite on the twigs. The cones are often 10 centimeters 
in diameter. The optimum development of the tree is variously reported at from 
350 to 1,500 meters above sea level. 

Distribution. — The species is found widely distributed throughout the Philip- 
pine Archipelago from northern Luzon to southern Mindanao. It is known to 
occur in the following islands and provinces: Lepanto, Zambales, Bataan, Cama- 
rines, Sorsogon, Negros, Palawan, Davao, and Zamboanga. 

Distinction from dammar resin. — Manila copal is frequently con- 
founded with true dammar. 

Ridley* speaks of "dammar minyak" as a large, coniferous tree {Dammara 
orientalis Lamb.) which grows in the hills of Penang, Perak, etc., of the Malay 
Peninsula. 

Wiesner,* in referring to this subject, says : 

During my sojourn in Java and Sumatra I had ample opportunity to see 
Dammara orientalis Lamb.* and its products and I can positively affirm that the 
resin is not identical with the dammar resin of commerce. The product of Dam- 
mara orientalis Lamb, compares with no known resin better than with the Kauri 
copal of Dammara attstralis Lamb. 

2 The Almaciga Tree. This Journal, Sec. A (1910), 5, 173. 
* Journ. Asiat. Soc. Straits, Burmah (1890), 92. 
^ Log. cit., 256. 

" Dammara orientalis Lamb, is treated by all authors as a synonym of Agathis 
alba (Lam). 



MANILA GOPAL. 179 

According to Tschirch," commercial dammar from the Malayan Archipelago i« 
very probably derived from dipterocarpous and not coniferous trees. 

As indicated by Livache/ Manila copal and dammar show widely different 
physical and chemical properties and (Jould not possibly be confused. 

IMPURITIES OF COMMEECIAL RESIN. 

The lack of uniformity, of the commercial product is largely responsible 
for the objection to Manila copal which exists among certain classes 
of consumers. Probably in no other industry does success more* largely 
depend upon the quality and uniformity of the ingredients than in. the 
manufacture of varnishes. The unscientific foundation upon which the 
art of varnish making is based has but to be realized to have appreciated 
the importance of this question of uniformity. Where opportunity has 
been at hand to note the primitive methods of collecting and grading 
varnish resins which prevail in this part of the world, it is easy to 
understand the difficulties to be overcome. 

At least four natural orders of resin-producing trees, namely, dip- 
terocarps, conifers, Burseraceoe, and Outtiferce, are found widely dis- 
tributed throughout the Malayan region, and their products are 
simultaneously collected, graded, and shipped to European and American 
markets. Therefore, it is not surprising that resins of widely varying 
properties and consequently of different values for a given purpose should 
become accidentally or intentionally mixed. 

Formerly the commerce in "Manila copal,^^ so far as concerns the 
Philippines, w^as largely in the hands of Chinese traders, who shipped 
directly to Borneo and Singapore, where the resin was undoubtedly 
largely mixed with dammar and similar products derived from these and 
other Malayan sources before being graded for reshipment. 

Even at the present time, 50 per cent of the* Philippine exportation 
of this commodity reaches the consumer via Singapore. The remaining 
50 per cent of the present local exportation is graded and sorted in 
Manila to meet the demands of American varnish manufacturers, to 
whom it is shipped directly ; but all collecting is done by wild hill tribes, 
and it frequently happens that a considerable admixture of dipterocarp 
resins is encountered by the Manila sorters. 

Hence it is seen that the varnish manufacturers must depend largely 
upon the different systems of sorting and grading as practiced by oriental 
shippers, and they in turn are dependent upon the indiscriminate mixing 
of good, bad, and indifferent varieties as practiced by the native col- 
lectors. The securing of a uniform quality or kind of varnish resin is, 
therefore, a matter of difficulty. 

°Die Harze und Harzbehalter. Leipzig (1906), 484. 
^Manufacture of Varnishes. London, 2d ed. (1908), 2, 83. 



180 RICHMOND. 



MODE OF PBODUOTION. 



Recent resin.— At least 50 per cent of the Manila copal exported from 
the Philippines at the present time is collected in the Davao district 
of Mindanao and probably 90 per cent of the resin produced in this 
region is obtained by blazing living trees. The best results are secured 
by removing strips of bark about 1 meter long and 20 to 30 centimeters 
wide from different sides of the tree, thus offering clean surfaces for 
the resin to deposit as it exudes from the cut edges of the bark. 

Another method of tapping practiced by the native collectors, which 
makes no provision for a clean surface upon which the resin may deposit 
and consequently gives a product which is generally mixed with chips 
and bark, consists in hacking a wedge-shaped place in the trunk of 
the tree. 

The resin first appears as almost colorless tears, which soon harden on 
their outer surfaces. As the exudation continues, the fresh resin, instead 
of flowing out over the first portion produced, appears to force the latter 
outward by depositing beneath it; that is, the outer surface is always 
hard and friable. Consequently, the hardening process is extremely 
slow, and the size of the lumps of hardened resin collected is dependent 
upon the length of time it is allowed to form. In this manner it requires 
about two weeks to produce pieces of solid resin of marketable size. 
No reliable information is at hand with respect to the quantity of recent 
resin which is produced by a single tree in a given period, or regarding 
the effect of seasonal changes upon the rate of production, although the 
native collectors state that the resin flows best just after the rainy season. 
The tapping or blazing of much longer areas of the tree trunk, as is 
practiced in Borneo, would undoubtedly give a much larger yield per tree. 



MINERAL RESIN. 



In distinction from the recent gum obtained as above described, the 
living trees have another way of producing commercial resin. When 
excavations are made so as to expose their roots, deposits of hardened 
copal possessing very similar chemical properties to the artificial resin 
are frequently encountered. The origin of such deposits is not clear, 
although it is well known that cutting or otherwise wounding the roots, 
as well as the stems, of living trees also induces a flow of resin. In New 
Caledonia the following method of producing kauri copal from different 
species of Agathis was formerly employed: Large pits were dug under 
the trees and cuts made on the exposed roots. The excavations were then 
carefully covered over and allowed to fill with resin, which hardened in one 
or two months into blocks. This method of production is also to some 
extent practiced in the Philippines and it furnishes a resin which at first 
is muddy in color and consistency, but which hardens in about two 



MANILA COPAL. 181 

months and then has much the appearance of the naturally formed 
mineral product. I can not explain the formation of resin not produced 
by intentional wounding of stems or roots of living trees. 

Miquel ® states that the resin of Dammara alha Lam. flows from the 
stems and collects on the roots in large lumps which are often washed 
away by streams and not infrequently are carried to the seashore in this 
manner. I have never seen large masses of resin on the stems or roots 
of an almaciga tree above the surface of the ground. Mr. Curran, of the 
Philippine Bureau of Forestry, reports that he has seen large stalactitic 
masses of hardened resin hanging from high, broken, or split branches of 
this tree. A continuous exudation and subsequent falling of resin from 
such a source offers a possible explanation for the occasional finding of 
considerable masses of resin on or below the surface of the ground near 
living trees. 

The theory has also been advanced that cracks in the roots may be 
caused by the pressure of accumulated resin. It seems very unreasonable 
to assume that the tree would as a physiologic process excrete a material 
so rich in hydrocarbons, which must play an important role in its life 
processes. On the other hand, it does seem very reasonable to believe 
that an injury from whatever cause to. the bark of branches, stems, 
or roots is the true explanation of any deposits of a resinous nature 
which may be found in the ground near the roots of living trees, [t 
is very conceivable that the bark of the roots is sensitive to serious 
attack from a number of sources, and it is undoubtedly true that the 
sticky; water-resisting nature of coniferous exudations in general affords 
a natural protection from the causes of vegetable decay. 

While present methods produce a limited quantity of very desirable 
mineral resin, they are extremely primitive and not adapted to general 
extension. The ground about the tree is first loosened with a holo or 
long heavy knife, and then scooped out with a coconut shell. When a 
sack full, about 3 arrobas (34.5 kilos), of crude resin is secured, it is 
carried down the mountains to the nearest seacoast town where it may be 
disposed of by sale or exchange. The uncertainty of the supply and the 
laborious and time-consuming nature of the work, makes this method 
of collecting mineral resin rather unprofitable 9,nd irregular. 

Occasionally, deposits of resins are found by digging out a decayed tree 
stump or by prospecting small openings in the forest, where surface 
indications suggest that one or more trees once stood. As for the pros-' 
pecting for resin where there are no surface indications of former forest 
growth, I am inclined to think that it is seldom if ever practiced. There 
are no denuded forest areas at the altitude of optimum development of the 
Philippine almaciga tree, such as occur in New Zealand where the most 

"Flora van Nederlandsich Indie. Supplement (1862), 1, 86. 



182 RICHMOND. 

valuable grades of fossil Kauri resin are found. That the Philippine 
forest regions do contain widely distributed deposits of mineral resin, 
irrespective of existing almaciga trees, is hardly to be questioned, but 
their isolated distribution and the extreme uncertainty of the work makes 
it very improbable that they will ever be extensively exploited. 

COMMERCIAL GRADING. 

All Manila copal exported from the Philippines directly to the United 
States is cleaned, sorted, and graded in Manila. In this regard, partic- 
ular attention is given to cleanliness, color, and size. 

Cleanliness is probably the most important value-determining factor. The 
resin, even with the most approved method of blazing trees and care in removing 
the hardened product, contains adhering or embedded fragments of bark, chips, 
and leaves, which are the despair of the varnish maker and which are very 
difficult to remove without materially diminishing the size of the pieces. On the 
other hand, the mineral resin is not nearly as liable to be contaminated with 
vegetable matter, and only requires to be chipped or scraped free of surface 
impurities of a mineral nature. 

Size. — Manila copal, as it comes to the graders, is extremely variable in respect 
to the size of individual pieces. This is due to its very friable character and the 
methods of collecting and handling, and the cleaning operations further increase 
the proportion of the smaller grades. 

Thus, there are shipped what are known as dust, pea, nut, thumb, and first 
or bold chunk sizes; and to prevent further division in subsequent handling and 
storage the resin is now shipped in wooden packing cases 76 by 44.5 by 45.70 
centimeters, and holding 90,718 kilos (200 pounds) net weight per case. 

Color. — The recent resin is always pale yellow in color. The mineral resin is 
usually reddish-brown, although it is sometimes found in light amber shades. 
Keeping in mind cleanliness, size, and color, several grades in market value are 
produced, ranging from No. 1 (first quality), consisting of light-colored, clean, 
bold pieces to the poorest quality consisting of dark-colored, dirty dust. 

Commercial uses. — The resin is almost exclusively used commercially in varnish 
manufacture, largely as an ingredient of oleoresinous varnishes, to a less extent 
as an ingredient of spirit varnishes. The solution of the resin in alcohol dries 
"flat" or opaque, hence it can not be employed. Manila copal also enters to some 
extent into the manufacture of sealing wax. 

Other uses to which it appears suitable are for the manufacture of cheap soaps 
and paper size. Like common colophony, it dissolves for the greater part in 
alkaline solutions, forming resinates of soda or potash which possess considerable 
detergent- properties. 

Aqueous solutions of the alkaline resinates are precipitated by solutions of 
all other metallic salts, e. g., aluminum sulphate, in the form of an insoluble 
resinate which could be used in paper manufacturing to render the paper non- 
bibulous. 

THE CHEMICAL EXAMINATION OF MANILA COPAL. 

Tschirch and Koch® examined two specimens, of resin {Dammara orientalis 
Lamb.) obtained from a firm of copal washers in Mainz, Germany, who guaranteed 
them to be genuine specimens of commercial Manila copal. The first specimen, 

^Arch. d. Pharm. (1902), 240, 202. 



MANILA COPAL. 183 

designated as soft, geographical origin not stated, consisted of nut-sized lumps 
of a yellowish-red color and completely soluble in alcohol. The second specimen, 
designated as hard, was from the Celebes and consisted of larger lumps than the 
first, clear yellow in color, and only partially soluble in alcohol. Both samples 
possessed practically analogous physical properties, but varied in their chemical 
constants as follows: 

I. II. 

Acid value 134 118 

Saponification value 190 165 

By applying Tschirch's method of the proximate analyses of resinous substances, 
which consists in dissolving the crude resin in ether and extracting successively 
with 1 per cent of ammonium, sodium, and potassium carbonates and subjecting 
the residue to steam distillation, they obtained the following results for the two 
samples in question: 



Soluble in sodium carbonate 



Soluble in ammonium 


I. 
Per cent. 

carbonate 4.0 


II. 
Per cent. 


I Insoluble in ammonium carbonate 


76.0 


80.0 


Etherial oil 




6.0 


6.0 


Residue 




12.0 


12.0 


Water 
Undetermined 




2.0 
1.0 


2.0 
1.0 



Insoluble in sodium carbonate - 



100.0 100.0 

From the above tabulated data it is noted that the hard and soft marked 
varieties of Manila copal have the same composition within the limits of experi- 
mental error of analysis for this class of substances, with the one exception that 
the soft, alcohol-soluble variety contained a small proportion (4 per cent) of acids 
which were soluble in a dilute solution of ammonium carbonate. 

Aside from numerous reports- on the physical and chemical constants 
of Manila copal which have appeared throughout the literature and in no 
instance of which are any botanical identifications given, the work of 
Tschirch and Koch constitutes all that is recorded concerning the con- 
stituents of "Manila copal.^' 

EXPERIMENTAL., 

The two marked varieties of "Manila copal/' namely, recent surface 
and mined resin, were obtained from the Manila Trading and Supply 
Company, the largest local graders and exporters of this commodity. The 
samples were taken directly from sacks of ungraded provincial shipments 
of known source. 

The surface resin came from Davao, Mindanao, and consisted of pale, lemon- 
yellow semitransparent pieces of nut and thmnb size, graded as "sorts." It 
was incompletely soluble in absolute alcohol, leaving a grayish, gelatinous, neutral 
residue which dried to a brittle resin. The crude resin was difficult to pulverize 
and possessed no well-defined melting point, but became somewhat softened and 
sticky at 100°. 

The acid value, determined by dissolving approximately 1 gram of the 
coarsely powdered resin in 50 cubic centimeters of absolute alcohol and 



184 RICHMOND. 

titrating directly with seminormal alcoholic potash in the presence of 
phenolphthalein gave the following numbers for a. mixed sample : 

Milligrams of potash 
for 1 gram of resin. 

N 
0.8704 gram rquired 4 cc. — potash 128.6 



2 
2 



0.9834 gram required 4.5 cc. — potash 128.1 

0.9762 gram required 4.45 cc. — potash 127.6 

1.1934 grams required 5.45 cc. — potash 128.1 



Average 128.1 

Individual pieces gave direct acid values varying from 120 to 130. 
Saponification values were determined by dissolving approximately 1 
gram of the resin in 50 cubic centimeters of absolute alcohol, then adding 
25 cubic centimeters of seminormal alcoholic potash and heating on a 
steam bath with a reflux condenser for one hour and titrating the 
excess of potash with seminormal alcoholic hydrochloric acid. The fol- 
lowing numbers were obtained. 

Milligrams of potash 
for 1 gram of resin. 

1.0452 grams required 6.66 cc. - potash 178.4 

1.1074 grams required 7.0 cc. — potash 176.9 

1.0622 grams required 6.66 cc. ^ potash 177.2 

0.9232 gram required 5.88 cc. ^ potash 178.3 

Average 177.8 

Somewhat higher saponification values were obtained by digesting 
with an excess of alcoholic potash for longer periods, and individual 
pieces of the resin showed variations of 20 units in the total amount of 
potash required when digested for a given period. 

The mineral resin. — The second variety was from the Island of Palawan. It 
consisted of brownish-yellow, irregular masses, varying from those of the first 
size to Ikrge, angular pieces of 1 to 2 kilos in weight. It was easily ground to a 
fine powder which was completely soluble in alcohol. Its direct acid and saponifi- 
cation numbers, taken from an average of four separate determinations of a mined 
sample, were 112.5 and 150.6, resspectively. Individual pieces of the mined resin 
gave direct acid numbers ranging from 100 to 120 and complete saponification 
numbers from 140 to 160. 

For the purpose of further comparison, four samples of resin of the 
following descriptions were selected from a sack of unsorted copal: 

1. Very brittle, nearly white, bold piece. 

2. Very brittle, dark brown, bold piece. 

3. Very pliable, nearly white, irregular mass. 

4. Very pliable, dark brown, irregular mass. 



Sample No. 

1 


Acid value 
milligrams potash. 

109.0 


2 


105.7 


3 


134.2 


4 


131.1 



MANILA COPAL. 185 

Direct acid and complete saponification determinations in duplicate 
gave the following numbers : 

Saponification value 
milligrams potash. 

148.0 

145.3' 

184.4 
181.2 

STEAM DISTILLATION OF THE CRUDE RESIN. 

One-half kilogram of coarsely powdered surface resin was subjected to 
steam distillation, but only traces of a volatile oil could be recovered. 
The mineral resin, which admitted of much finer pulverization, yielded a 
larger portion of its content of volatile oil. 

A large, irregular mass of the very pliable resin (samples N"os. 3 and 4 
above) also yielded a trace of oil to exhaustive steam distillation. 

Upon the addition of alkali in slight excess of the amount required to neutral- 
ize the resin, and continuing the distillation, from 500 grams of surface resin 
(sorts) there were obtained 6.5 grams of oil or 1.3 per cent; from mined resiu 
(bold pieces), 40.0 grams of oil, or 8 per cent; from soft resin (sample No. 3), 
56.0 grams of oil, or 11.2 per cent. 

It is seen from the above that the state of aggregation rather than the age 
of the resin influences the proportion of volatile oil which can be isolated 
by steam distillation ;;the soft, pliable product and the brittle, bold piece of 
mined copal containing a much larger proportion of volatile terpenes than 
did the small, friable pieces obtained from living trees by artificial means. 
Examination of the volatile oil for possible differences due to the above 
sources was not made. The combined fractions, dried over solid potash, 
gave a pale, lemon colored oil of very pleasant odor. Specific gravity 

i^=0.865; n5^- = 1.4648 ; A-- = -26.55. The oil was distilled over 

sodium, the main fraction boiling from 155° to 165°, leaving about 10 
per cent by volume above 165°. 

The fraction from 155° to 165° was converted into a hydrochloride 
which recrystallized from alcohol in needles melting at 124° and which 
was identified as pinene hydrochloride. For further examination of 
the terpenes from Manila copal see Brooks, This Journal, Sec, A 
(1910), 5, 205. 

The main objection to Tschirch's method of examining resinous sub- 
stances is its inapplicability to resins which are not completely soluble in 
ether or some other solvent indifferent to aqueous alkalies. Furthermore, 
the difficulty of complete removal and separation of the weak resin acids 
from their etherial solutions with dilute solutions of ammonium and 
sodium carbonates renders the method extremely tedious and unsatis- 
factory. 



186 RICHMOND. 

In the case of Manila copal, which is only partially soluble in ether, 
a much better scheme of analysis was found to be as follows : 

The crude, mined resin is dissolved in strong alcohol in which it is completely 
soluble except for the varying proportion of dirt and other foreign matter which 
may be present. The filtered solution is then made alkaline by the addition of 
alcoholic potash in slight excess, which precipitates a white, semigranular mass 
which adheres to the bottom of the flask and from which the clear, supernatent 
alcohol may readily be decanted. The alcohol is then removed by distillation 
and the residue taken up with water and extracted with ether. The etherial 
extract is dried over solid potash and evaporated to constant weight. The residue 
is a pale lemon colored, mobile oil of terpene-like odor. 

The aqueous portion is warmed to expel the dissolved ether and acidified with 
dilute hydrochloric acid in slight excess, the precipitated resin washed free from 
potassium chloride and dried under reduced pressure over sulphuric acid to 
constant weight. 

The semigranular resinous mass which is precipitated from the alcoholic 
solution of the original resin upon neutralization with alcoholic potash is treated 
with hot water, in which it is soluble, except for a small proportion of gelatinous 
residue which can not be completely separated by filtration. Extraction with 
ether in which the insoluble portion is partially soluble effects a complete me- 
chanical separation from the soluble alkaline resinate. The ether extract con- 
taining the partly dissolved, unsaponified portion is evaporated to dryness at 
100", leaving a grayish, brittle, amorphous resin. 

The aqueous portion of the precipitate, insoluble in alcohol, is in turn washed 
to expel dissolved ether and the resin precipitated with dilute hydrochloric acid. 
The washed, amorphous resin is dried to constant weight. 

One hundred parts of the crude resin treated in this manner gave the 
following : 

Parts. 

Insoluble in absolute alcohol 0.5 

Soluble in alcoholic potash 40.0 

Insoluble in alcoholic potash 41.5 

Neutral oil soluble in alcohol and volatile with steam 6.0 
Neutral resin partially soluble in alcohol and nonvolatile 

with steam 10.0 

Water, etc., undetermined 2.0 



100.5 



These results agree very closely with those reported by Tschirch for a 
sample of spirit soluble Manila copal, and confirm his general conclusions : 

(1) That Manila copal consists mainly of amorphous, free resin acids; 

(2) it contains a neutral resin, indifferent to alkalies; and (3) a volatile 
oil 

However, it should be noted that Tschirch's so-called hard variety 
contained no acid extractable with ammonium carbonate and was only 
partially soluble in alcohol; whereas the hard mineral resin examined in 
this laboratory is soluble in alcohol and does contain a small proportion of 
an acid soluble in ammonium carbonate. 



MANILA COPAL. 187 

THE RESIN ACIDS. 

Over 80 per cent of the crude resin is soluble in dilute aqueous solu- 
tions of the fixed alkalies and is precipitated as a pale yellow, amorphous 
resinous solid when neutralized. Like the original resin, it has no well- 
defined melting point, but begins to soften at 100° and fuses to a clear, 
viscous, amber-colored resin with slight decomposition at 150°. It is 
completely soluble in ethyl alcohol and phenol, partly soluble in ether, 
benzene, and chloroform, and very slightly soluble in petroleum ether and 
ligroin. Its solution in absolute alcohol is partly precipitated with 
alcoholic potash or alcoholic lead acetate in slight excess. 

The alcohol soluble and insoluble potash salts were approximately 
separated in about equal proportions by decantation of the neutral alco- 
holic solution. The latter was evaporated, the residue dissolved in water, 
and the free resin acid liberated with dilute hydrochloric acid. The resin 
acid thus obtained could not be crystallized. It began to soften at about 
70° and \vas quite fluid at 100°. It was completely soluble in ether 
and contained an acid soluble in a dilute solution of ammonium carbonate. 

Tschirch and Koch ^^ obtained a crystalline acid by extractino^ an ether solution 
of soft Manila copal with a 1 per cent solution of ammonium carbonate, to vvliich 
he assigned the formula CgHioOj, based upon elementary analyses, molecular weight 
determinations, and analyses of its potassium salt. The acid crystallized from a 
mixture of ethyl and methyl alcohol in needles. Its mjelting point was 175" 
and a 2 per cent alcoholic solution gave a rotation of 2® 24'. 

From Kauri copal {Dammara australis) he obtained*^ a crystalline acid in 
the same manner which melted at 192° and which gave a rotation of 2° 24'. Ks 
analysis and melocular weight corresponds to the composition doHioOa. 

To obviate the tediousness of repeated extractions with 1 per cent 
ammonium carbonate which Tschirch experienced, the etherial solution 
of the resin acid was shaken with a large excess of 5 per cent ammonium 
carbonate solution in a 10-liter bottle on a motor-driven shaking machine 
for several days. A vent through the stopper of the bottle allowed of 
equalization of the pressure. Complete extraction gave 4 per cent of 
an acid calculated on the amount of the original resin. It was precip- 
itated in an amorphous form when the ammonium carbonate solution 
was acidified. It was dissolved in dilute alcohol, from which it crystal- 
lized upon long standing in the cold in needles. The recrystallized acid 
was perfectly white and melted at 185° to 187°. In addition to the 
solvents of the crude resin acids, the crystalline acid is also completely 
soluble in ether, benzene, methyl alcohol, and chloroform. 

^^Loc. cit., 209. 

^^Arch. d. Pharm. (1901), 239, 152. 
95193 2 ; 



188 RICHMOND. 

Combustions*- of the substance dried at 100° gave the following results: 

I. 0.1820 gram substance gave 0.4783 gram C^O, and 0.1487 gram H,0. 
II. 0.2363 gram substance gave 0.6202 gram CO, and 0.1823 gram H.O. 

III. 0.3261 gram substance gave 0.8509 gram CO, and 0.2510 gram H,0. 

IV. 0.2292 gram substance gave 0.6029 gram CO, and 0.1866 gram k'o. 
V. 0.2058 gram substance gave 0.5419 gram CO, and 0.1700 gram HA 



Calculated 
for C10H15O2. 
Per cent. 
C=71.85 


I. 
71.66 


Found 'per cent. 

n. III. IV. 
71.58 71.66 71.72 


V. 

71.85 


H= 8.98 


9.06 


9.13 9.00 9.03 


9.17 



Cryoscopic determination of the crystalline acid in glacial acetic acid gave 
molecular weights as follows: 

Weight of Weight of Depression of Calculated 

solvent. substance. freezing point, molecular weight. 

25.9380 0.171 0.150 170.5 

25.949.6 0.138 0.118 174.9 

25.4211 0.1686 0.150 171.5 

25.4211 0.093 0.086 165.0 



Average 170.5 

The basicity was determined by titration of an alcoliolic solution of the 

N 
acid with — alcoholic potash in the presence of phenolphthalein. 

Milligrams of potash 
for 1 gram of acid. 

N 
0.5 gram required 29.5 cc. — potash 33 1 

N 
0.6745 gram required 39.5 cc. - potash 332 

therefore 56.1 grams of potash would neutralize 1()9 grams of acid. The acid is 
therefore monobasic. A 2 per cent alcoholic solution in a 20 cubic centimeter 
tube gave a rotation of -f 2.7°. 

The residue, after complete extraction with ammonium carbonate to 
separate the crystalline acid, was converted into its potassium salt in the 
usual manner. Every conceivable means to induce crystallization was 
tried, but without success. The potassim and sodium salts appear as 
microscopic needles upon the addition of solid potash or soda to an 
aqueous solution of the resinate, but they were too hygroscopic to allow 
of complete purification. The alkaline earth and heavy metal salts of the 
acid are amorphous. 

"The elementary analyses of resin acids recorded in this paper were made by 
Mr. Mariano Vivencio del Rosario, assistant professor of chemistry in the Philip- 
pine Medical School. 



MANILA COPAL. 189 

Combustions of the noncrystalline acid, rendered as pure as possible through 
its potassium salt, gave the following results. 

I. 0.2190 gram substance gave 0.5868 gram CO2 and 0.1813 ^ram HA 
II. 0.2072 gram substance gave 0.5559 gram CO, and 0.1681 gram H2O. 

Calculated for C22H34O4. Found per cent. 

Per cent. I. 11. 

C = 72.92 73.01 73.13 

H= 9.39 9.19 9.07 
Molecular weight: 

Weight of solvent Weight of Molecular weight 

acetic acid. substance. Lowering. calculated. 

25.4961 0.2134 0.075 377 

25.4975 0.1302 0.056 353.8 

25.3034 0.1445 0.061 360. 

The determination of the basicity gave the following numl)ers: 

Milligrams of potash 
for 1 gram of acid. 
N 
1.0466 grams substance required 28.9 cc. r-: alcoholic potash 156.3 ' 

N 
1.0113 grams substance required 30.4 cc. -- alcoholic potash 155.6 

whence 56.1 grams of potash would neutralize 361 grams. Therefore the acid 
is monobasic. 

DISTILLATION UNDER REDUCKD PRESSURE. 

One attempt at distilling the low-melting, amorphous acid was made 
under a pressure of 3 to 5 millimeters; 75 per cent came over between 
240° and 275° with but slight decomposition. The clear, amber-colored 
distillate was dissolved in aqueous potash and extracted with ether which 
removed a small amount of oily hydrocarbon. The acid recovered from 
its potassium salt possessed the same properties as before distillation; 
with a pressure of 1 millimeter or less it may be possible to distill 
this acid entirely unchanged, and this offers the most promising method 
of getting the substance sufficiently pure for a study of its constitution. 

RESIN INSOLUBLE IN ALCOHOLIC POTASH. 

The semigranular mass which was thrown out of the alcoholic solution 
of the original crude resin, upon the addition of alcoholic potash in 
slight excess consisted of a potassium salt soluble in water, a «ieutral 
residue insoluble in water, but soluble in boiling aqueous potash, and 
a small proportion of unsaponifiable matter. 

The w^ater soluble portion was precipitated as a. white, amorphous 
powder upon acidification with dilute hydrochloric acid. It was com- 
pletely soltible in ethyl and amyl alcohol and aniline and slightly soluble 



190 RICHMOND. 

in ether, benzene, and acetic acid. It did not fuse under 220° and con- 
tained no substance which could be extracted with ammonium carbonate. 
Combustion of the dry, noncrystalline substance gave the following figures: 

I. 0.1817 gram substance gave 0.5150 gram CO2 and 0.1634 gram HjO. 
II. 0.2114 gram substance gave 0.5970 gram CO2 and 0.1912 gram H^O. 

Calculated for C32H50O4. Found per cent. 
Per cent. I. II. 

C=77.11 * 77.27 77.05 

H= 10.04 9.98 10.02 

Molecular weight =498. 
The determination of the basicity was as follows: 

Milligrams of potash 
per gram of resin. 
N 
1.0045 grams required 20.4 cc. —potash 113.9 

N 
1.0567 grams required 21.0 cc. — potash 112.0 

whence 66.1 grams potash would neutralize 496 grams of acid. It is therefore 
monobasic. 

That this last substance was manifestly impure was subsequently 
shown by the fact that the precipitate obtained by neutralizing an 
alcoholic solution of* the dry substance with alcoholic potash was never 
entirely soluble in water. All known potassium salts of resin acids 
dissolve readily in water. The insoluble residue was also indifferent to 
hot aqueous potash, which precludes lactone formation. 

The insoluble resin which was always left behind by this means 
possessed all the properties of the neutral resin which was obtained in 
the same manner from the original, crude copal. It was entirely inert 
to strong alkalis and scarcely soluble in the ordinary organic solvents, 
but entirely soluble in an alcoholic solution of the resin acids. 

In view of the foregoing, no particular importance is placed upon the 
results of the elementary analyses of the substance, or the close agree- 
ment of the molecular weight with that of a monobasic acid of the 
formula C^jgHgoO^ as determined from the potash value. 

NEUTRAL RESIN. 

A sHiall portion of the resin which was precipitated from the alcoholic 
solution of the original copal when the latter was neutralized with 
alcoholic pgtash was also insoluble in water, slightly soluble in alcohol 
and ether, but completely soluble in alcoholic solutions of the resin acids. 
It is partially saponified by hot digestion with aqueous potash, leaving 
a small proportion of substance practically insoluble in all solvents and 
perfectly inert to alkalies. 

The saponification value of this neutral resin was found to agree 
with the so-called ester number of the crude resin, and the amorphous 
acid obtained from such saponification possessed properties analogous to 



MANILA OOPAL. 191 

those of the impure, high-melting resin acid obtained from the portion of 
the original resin insoluble in alcoholic potash. 

The unsaponifiable matter does not melt, but chars at once upqn the 
application of direct heat, hence further examination of it was not made. 

Too much emphasis can not be placed upon the difficulties encountered 
in a chemical examination of this class of substances. They are as 
follows : First, their great susceptibility to atmospheric oxidation as 
shown by Brooks ^^ demands the avoidance of heat and. undue exposure 
to air and light; second, the tenacity with which the amorphous sub- 
stances precipitated from aqueous solutions retain moisture when dried 
under the imposed conditions; third, the possibility of isomeric forms of 
the same substance; fourth, the tendency of resin acids to lactone or 
other anhydride formation under the influence of heat and chemical 
reagents; and fifth, the conversion of resin acids to neutral, oxygenated 
bodies accompanied with a gradual lowering in potash value due to los^ 
of carbon dioxide. 

SUMMARY. 

It should be repeated that the experimental results obtained are not 
considered sufficiently definite or extensive to warrant more than the 
general conclusions that Manila copal consists essentially of free amor- 
phous acids, a volatile hydrocarbon, a neutral saponifiable substance, 
probabl}^ a lactone, and an unsaponifiable resin. 

Contrary to expectation, the free acids appear to bear no relation to 
each other, or to the known resin acids of other coniferous resins. 

It should also be pointed out that the observations concerning the 
resin acids of Manila copal herein recorded do not confirm those of 
Tschirch and Koch for the resin acids of Dammara orientalis Lamb. 

OLEOKESINOUS VARNISHES. 

In distinction from spirit varnishes, which are solutions of resin in 
some volatile solvent such as turpentine, benzene, alcohol, etc., and from 
plain oil varnishes which consist only of linseed or some other drying oil, 
oleoresinous varnishes contain all the ingredients and partake of the 
properties of both spirit and oil varnishes. 

The art of making oleoresinous varnishes has undergone no important 
changes since the days of its first application. It consists essentially 
now, as it did four hundred years ago, in combining resin, linseed oil, and 
turpentine in such a manner as to produce a clear, homogeneous liquid. 
Naturally, the simplest method of preparing such a varnish would be to 
dissolve the resin in the drying oil in the right proportions to give the 
desired effects of hardness and elasticity, and then to dilute the solution 
with turpentine or some other volatile solvent to a consistency suitable 

^^This Journal, Sec. A (1910), 5, 219. 



192 RICHMOND. 

for use. However^ it has been found in commercial practice that natural 
resins will not dissolve in raw drying oils. In some instances, a fairly 
clear solution is produced by a little heat^ but ' almost invariably the 
dissolved resin will either separate from the oil upon cooling, or it will 
be precipitated upon the addition of turpentine; hence, the practice of 
subjecting the oil and the resin, both before and after mixing, to elevated 
temperatures is almost universal. 

A brief description of the present practice of making oleoresinous 
varnishes seems necessary. The following is the method according to 
Sabin :^* 

Melting the resin. — The American practice is to lieat about 125 pounds of resin 
at a time in a copper kettle over a coke fire. When all the lumps are melted and 
the resin is quite liquid, i. e., when it drops from the stirring rod, it is removed 
from the fire. By this time, which requires about half an hour, from 10 to 20 
per cent of the resin is lost in the form of pungent, irritating, higlily inflammable 
vapors. A thermometer is not commonly used, for the temperature is not so 
essential as the melted feel imparted through the stirring rod. The temperature 
is seldom below 343° C. when the melting is completed. The temperature and 
the percentage of loss vary greatly with different resins. 

Adding the oil. — When the resin is all melted and removed from the fire, the 
linseed oil, made ready in another kettle, is slowly added with constant stirring. 
The oil is previously heated to about 260° C. Of course, if only a little oil is 
to be added the temperature does not have much effect on the mass, but it is 
common to have the oil hot. 

Cooking the varnish. — As soon as the oil is added the kettle is put back on 
the fire. Although the mixture appears to be a complete solution, it is not really 
so at this stage, for if a drop of it be removed and allowed to cool the resin will 
separate, making the drop cloudy, and the common rule is to withdraw the 
stirring rod from time to time and let a drop of the mixture fall on a piece of 
glass, where it cools at once and shows by its cloudiness that combination has not 
or, by its clearness, that it has taken place. A more approved practice is to keep 
a thermometer in the liquid and heat to a certain temperature for a certain length 
of time, previously determined as the best for the particular varnish which is 
being made. This temperature, roughly speaking, is not very far from 260°. 

Thinning down tvith turpentine. — When the oil and resin have been properly 
cooked, the kettle is withdrawn from the fire and allowed to cool somewhat and 
the liquid diluted with turpentine to standard consistency. 

It will be seen from the above description of American varnish manu- 
facture, which in its main features holds true elsewhere, that the art of 
making oleoresinous varnishes rests entirely upon an empirical basis; 
and that the results of practical experience in the art of treating and 
blending the unmi-xed ingredients have been handed down through 
generations to the present time without particular regard for the un- 
derlying causes of the effects thus produced. 

There is a steadily increasing demand for resins of the type known 
as copal^ which when dissolved in drying oils produce varnishes distin- 
guished by leaving a hard, lustrous, transparent coating when applied to 

^* Technology of Paint and Varnish. New York (1907), 71-81. 



MANILA COPAL. 193 

surfaces. As is well known, ^'Manila copal'^ is employed almost ex- 
clusively in the varnish industry and together with Kauri copal, a 
closely related product of the same genus of tree, enjoys the most 
extensive use of all varnish resins. 

In common with all other resins employed for the same purpose, 
^^Manila copaF^ is almost invariably heated to such a temperature and for 
such a period as to render it perfectly liquid. Whether the operation is 
termed roasting, fusing, melting, or running, the reason is the same, 
namely, to render the resin miscible with linseed oil. Its behavior on 
fusion is therefore, from a practical standpoint, the most important prop- 
erty to determine. 

ACTION OF HEAT UPON "MANILA COPAL." 

Continued roasting of the powdered resin at 100° causes a slight 
darkening in color. 

Tlie recent surface resin becomes somewhat softened, but older, mined 
specimens are not affected at this temperature. Both varieties show 
a very appreciable gain in weight on continued exposure at 100°. One 
gram of finely powdered mined resin gained 4 per cent in weight in 
seventy-two hours. Coarsely powdered surface resin showed a slightl}^ 
less increase. As both samples contained water and volatile hydrocarbons, 
a gain in weight under the above conditions is significant of quite pro- 
nounced oxidation. 

One hundred grams of coarsely powdered copal were placed in a distilling 
flask and heated continuously for forty-eight hours at a temperature of 100°. 
Approximately 2 grams of water vapor were condensed. Instead of the equivalent 
loss in weight the residue showed a slight gain. Coarsely powdered resin ift a 
desiccator over sulphuric acid, either at ordinary pressures or in a partial 
vacuum, quickly discolors the acid, which in a few days becomes wine red, showing 
that volatile organic matter is easily liberated under these conditions. 

DRY DISTILLATION OF MANILA COPAL. 

Five hundred grams of coarsely powdered resin were heated in a large 
flask connected with a long condensing tube. Some water and aromatic 
oil was condensed between 100° and 200° as registered by a thermometer 
in the vapor. 

Considerable care in heating was necessary to prevent the swollen, 
semifused mass from passing over into the condenser. Finally, the 
mass subsided and boiled gently with an occasional liberation of gray, 
uncondensible vapor. Three hundred and fifty grams of greenish-yellow 
oil were collected below 350°, leaving a pitch-like residue, amounting 
to 30 per cent of the original copal. 

The oily distillate was fractioned at ordinary pressures without appreciable 
decomposition. 

First fraction, 140° to 200°; 122 grams of reddish ail containing a little water. 



194 RICHMOND. 

Second fraction, 200° to 250°; 13.3 grams of reddish oil which contained no 
water. 

Third fraction, 250° to 300° ; 35.5 grams of greenish, mobile oil. 
Fourth fraction, 300° to 350°; 191 grams green oil. 
Residue, above 350°; 98 grams by difference. 

Nearly 65 per cent of the crude distillate boiled between 250° and 
350°. The properties of this main fraction showed a general agreement 
with those of the known diterpenes and it is regarded as belonging to 
this class of compounds. ^^ 

CHANGES WHICH TAKE PLACE WHEN "MANILA CQPAL" IS FUSED IN AN OPEN 
RECEPTACLE OVER A FREE FLAME. 

The surfaces of the pieces first soften and adhere to each other and to 
the sides of the kettle. As the heating is continued the resin becomes 
less viscous and sticky and can readily be stirred^ and heavy^ gray vapors 
of a strongly acid reaction are thrown off. During the disengagement of 
the steam and acid fumes the semifluid mass froths and shows a tendency 
to rise over the sides of the kettle unless constantly stirred^ and some- 
times the removal of the flame is necessary to prevent loss in this manner. 
Eventually, the fused mass subsides to a clear, dark, amber-colored 
mobile fluid which boils gently at a temperature between 300° and 350°. 
Upon cooling, the solution quickly becomes viscous and finally sets to 
a transparent, glassy mass which may readily be powdered. ' 

The copper kettles employed in varnish factories for melting Manila 
copal become coated with a greenish incrustation which is undoubtedly 
due to the action of the acid fumes liberated during the melting process. 

That profound chemical and physical changes take place when organic 
matter like resins and vegetable oils are subjected to elevated temper- 
atures is very manifest. • 

A study of the finished product offers almost insurmountable difficulties in 
the way of determining its composition, for the varnish is a mixture of complex 
ingredients rendered still more complex by the temperature to which they are 
subsequently subjected. No better illustration of this fact can be cited than the 
results obtained in this laboratory on a sample of varnish manufactured by a 
well-known and reliable firm and claimed by them to consist only of resin and 
vegetable drying oil and containing no resin oil whatever. Exhaustive examina- 
tion of the varnish in question showed conclusively the presence of resin oil, 
which, if not added directly, is undoubtedly a product of the process of manu- 
facture. 

It would seem that a much better method of attacking the problem 
of what an oil varnish consists of and the changes which the raw materials 
undergo during the process of its manufacture would be to study the 
effect of heat upon the unmixed ingredients ; for, aside from the possible 
difference in the effect of heating the oil and the resin for the same 

" For a detailed study of the dry-distillation products of Manila copal see 
Brooks, This Journal, Sec. A, 5, 203. 



Raw copal. 


Fused copal. 


127.6 


68.21 


175.17 


136.3 


137.9 


133.3 


15.98 


22.99 



MANILA COPAL. 195 

period at a given temperature separately, or in admixture, the results 
should be comparable and at the same time the problem would be very 
much simplified. 

Williams *" examined samples of Kauri and copal not only in their original 
condition, but also after being subjected to the melting or "running" process, 
which is used for converting them into varnish. He observed an appreciable 
decrease in the acid, saponification and iodine values upon melting. 

Lewkowitsch " determined the constants of the more common varnish resins both 
before and after heating them to 300°. For a reputed specimen of Manila copal 
he obtained the following results: 

Acid value 
Saponification value 
Iodine value 
Per cent unsaponifiable 

Lewkowitsch also observed that the fused resin gained in weight upon heating 
for fifteen hours at 100° and noted that this fact indicates oxidation. 

It was thought that further study of the changes which take place 
when Manila copal is heated at different temperatures and for periods 
corresponding to the heating which it subsequently receives in the cooking 
stage after the oil is added would be of value. 

Experiment 1. — One hundred and twenty grams of nut and thumb sized pieces 
of mined resin were heated in a porcelain casserole over a free flame with constant 
stirring with a thermometer until the temperature of the mass reached 260°. 
At this point the mass was not entirely homogeneous and did not drop freely 
from the stirring rod. Upon cooling it set to an amber-colored, brittle solid 
filled with gas bubbles. It weighed 104 grams, a loss of 13.3 per cent. 

Experiment 2. — One hundred and twenty grams of the resin were heated Until 
the temperature registered 275°, when the melted mass was a homogeneous, 
mobile liquid which dripped from the stirring rod in free, oily drops. The residue 
weighed 101.5 grams, a loss of' 15.4 per cent. 

Experiment 3. — One hundred and twenty-five grams of resin were heated up 
to 300° and the cooled mass weighed 105 grams, a loss of 16 per cent. . 

Experiment Jf. — One hundred and twenty-five grams of resin were heated to 
325° and the cooled mass weighed 103.3 grams, a loss of 17.4 per cent. 

The residues from experiments 2 and 3 were separated into their constituents 
exactly in the same manner as the raw resin was treated, namely, dissolved in 
alcohol, filtered, neutralized with alcoholic potash, and the aqueous solutions of 
the alcohol soluble and insoluble potassium salts extracted with ether. 

From 100 parts by weight of the fused residues there were obtained 
the following: 

Experiment 2. Experiment 3. 
Dirt and foreign matter insoluble in alcohol 0.36 0.4 

Amorphous resin acids the potassium salts of 

which were soluble in alcohol 42.00 39.5 

Amorphous resin acids the potassium salts of 

which were insoluble in alcohol 
Neutral oil soluble in ether 
Neutral resin by diff'erence 

^'Analyst (1898), ,23, 254. 
'' Loc. cit. (1901), 26, 37. 



53.40 


52.7 


0.00 


2.5 


4.24 


4.9 



196 RICHMOND. V 

The samples of resin for the above experiments and for that which 
was examined in the imfused state (see p. 183) were taken from a 
large piece of mined resin weighing about 2 kilograms, hence any changes 
in the recorded constants • are due to the effect of heat under the con- 
ditions given above. 

The saponifiable portion, that is, the free, amorphous acids, were united and 
again heated in the same manner as in the previous fusions to their respective 
temperatures, namely, 275° and 300°, and held at those temperatures with 
frequent stirring for one hour each. Both showed an appreciable loss in w^eight 
and the consequent formation of neutral resin oil. 

While resin oil is miscible in all preparations with linseed oil and is 
considered a good solvent for resin acids, and although it undoubtedly 
does play some part in holding resin acids in solution in boiled oil, it is 
not believed that its formation in such relatively small amounts has any 
important role in the manufacture of oleoresinous varnishes. 

After the second fusion of the united resin acids obtained from experiments 
2 and 3, the masses were dissolved in alcohol, neutralized with alcoholic potash, 
and separated into alcohol soluble and insoluble potassium resinates as before. 
The liberated resin acids behaved in a manner entirely similar to those obtained 
from the crude resin. They could not be made to crystallize, hence combustions 
and cryoscopic determinations 'were not repeated, but their melting points, solu- 
bilities, and other physical properties showed no appreciable alteration. 

The fused acid gave an acid value of 164 milligrams of KOH, as compared with 
169 to 170 before fusion. 

Differences in the behavior of the fused and unfused resin acids indica- 
tive of profound decomposition or depolymerization were not observed. 

It has been shown that the melted resin differs from the raw resin only 
in the amount and nature of the unsaponifiable portion, and that the resin 
which enters into varnish manufacture consists essentially of free acids 
of the same composition as they had when in the original copal. There- 
fare, it would seem that any real or assumed reason for subjecting the 
oil and resin to elevated temperatures, both before and after mixing, 
should be looked for elsewhere. 

THE USE OF MANILA RESIN FOR OLEORESINOUS VARNISHES WITHOUT PREVIOUS 

FUSION. 

Owing to the loss of weight and the considerable darkening in color 
which all resins undergo during the melting process, many attempts have 
been made to render raw copals soluble in linseed oil. 

Livache** makes turpentine varnishes from unmelted Kauri and Manila copals 
by first dissolving them in alcohol to which a trace of nitric acid has been added, 
then adding the turpentine and removing the alcohol by distillation. To make 
oleoresinous varnishes, the spirit varnishes thus prepared are dissolved in a 
mixture of linseed oil and the free fatty acids of linseed oil. 

^^Gompt. rend, Acad. sci. (1907), 146, 896. 



MANILA COPAL. 197 

Andes " also refers to the use of linseed-oil acids in giving elacticity to spirit 
varnishes and as a solvent for copals. 

Acting upon the above suggestions and from tlie knowledge that 
substances of an acid nature such as alcohols, phenols, acetic acid, etc., are 
much more efficient solvents for resin acids than neutral bodies like 
petroleum ether, benzene, or esters, it was found that oleic, palmitic, and 
linolic acids, either alone or in admixture, dissolved the resin acids of 
Manila copal at moderate temperatures to clear, pale, amber-colored 
solutions which remain so in the cold. Hence, an inquiry into the 
nature of varnish oils and the changes which they undergo in the man- 
ufacture of oleoresinous varnishes was considered important in its bearing 
upon the subject under discussion. 

The first function of a vegetable oil as an ingredient of an oleoresinous 
varnish is to serve as a vehicle or solvent to bring the resin into a 
fluid condition and maintain it there. 

Any vegetable oil, such as coconut or cotton-seed oil, will fulfill this 
function, but a solution of resin in these oils would never dry. The 
second function of the vegetable oil is, therefore, to provide a solvent 
which will dry and thus bind the particles of dissolved resin together in 
a thin, elastic film; hence the so-called drying oils of which linseed oil 
is the type have come into almost universal use for this purpose. 

Dismissing the various theories which have been advanced as to "why 
and how oils dry,^^ with the statement that because of their unsaturated 
nature they are capable of absorbing large quantities of oxygen and that 
by various means they may be rendered more siccative or drying, attention 
will be called to the different preliminary treatments of linseed oil for 
use in the arts and trades. 

According to Andes,-^ boiled linseed oil produced at any temperature below 
230° can be used only for the grinding and thinning of paints or for application 
alone. The varnish manufacturer requires boiled oil which has been heated to 
at least 270°, so that all the foreign matters contained in the oil have been 
removed. A difference must therefore be made between boiled oil intended for 
paint and for varnish making. By foreign matters, Andes refers to muc^age and 
the yellow pigment to which the color of raw linseed oil is due. 

Weger ^^ makes the following remarks concerning linseed-oil "mucilage:" In 
the manufacture of oil varnishes "yarnish linseed oil'* is used. By this term is 
understood an oil which when rapidly heated to 280° to 320° does not "break," 
i. e., does not separate any solid substances. If fresh, raw linseed oil be tested 
by rapidly heating it in a test tube, a turbidity generally appears at 250*. 
Ordinarily, commercial linseed oils (with the exception of "varnish oil," which 
has undergone some chemical or mechanical treatment) almost always show this 
coagulation upon rapid heating. 

^* Drying Oils, Boiled Oil, and Solid and Liquid Driers. London (1901), 68. 
^ hoc. dt., 174. 
" hoc. cit., 125. 



198 RICHMOND. 

Fehr" considers that the loss of mucilage is the most important result of 
boiling oil, as it retards the drying, or more properly the hardening of the varnish 
film. 

Mulder,*^ on the other hand, found no difference in the drying properties of 
raw linseed oil which had been filtered through charcoal, which must remove even 
dissolved mucilage, since the filtered oil did not "break" on heating. 

This foreign matter designated as mucilage or albumin is present 
in all expressed vegetable oils, but in such relatively small quantities 
that it seems unreasonable to think its removal to be the main reason 
for the specific treatment to which all raw oil is subjected in the prepara- 
tion of boiled and varnish oils. 

Mulder 2* concluded that linseed oil which had only reached the first stage 
in its oxidation contained free fatty acids. Bauer and Hazure ^^ made the same 
observation: "A sample of linseed oil four years old which was covered with a 
thin skin, but otherwise completely soluble in ether, contained 8.9 per cent of 
free acids. Another sample which had been stored in a badly closed flask for 
five years contained 12.2 per cent of free acids." Lewkowitsch 2« gives the free 
acid values calculated as oleic acid in weakly and strongly oxidized linseed oil 
as 18 and 28.9 per cent, respectively. The acid values of some commercial varnish 
oils recorded by Fahrion range from 13.4 to 32.6 per cent. According to Sadtler,='^ 
when linseed oil is boiled so as to have lost about 8 per cent of its weight, it is 
converted into ordinary boiled-oil varnish, and if it is heated until it has lost 
about 16 per cent of its weight it becomes .t^iicker and yields a stiff varnish used 
as the basis for printing inks. 

Linseed oil contains from 9 to 10 per cent of glycerol, which cor- 
responds approximately to the loss in weight which the oil suffers when^ 
converted into ordinary boiled-oil varnish, but boiled linseed oil also 
contains considerable quantities of glycerol; hence other changes involv- 
ing the formation of volatile products besides the hydrolysis of the 
glycerides and the loss of the glycerol or ■ its decomposition products, 
acrolein and water, undoubtedly take place. The highly inflammable 
nature of the gases which are given off when linseed oil is heated indicate 
hydrocarbons and suggest more profound decompositions than mere 
hydrolysis. However, linseed oil is not a very stable compound, con- 
sisting as it does of weak acids in combination with a weak base ; and 
whether the varnish oil is prepared by heat alone, or in conjunction Vith 
added driers such as the oxides of manganese or lead, which are much 
stronger bases than glycerol, considerable dissociation must necessarily 
take place. In the case of the preparation of lithographic varnishes and 
printing inks, where the chief requirement is a product which will not 

^^Loc, cit., 150. 

^'Loc. cit., 126, 

=^* Loc. cit.^ 8. . 

'''Monatsh. f. Chem, (1888), 9, 469-468. 

2«0ele, Fette u. Wachs (1905), 2, 592. 

^Industrial Organic Chemistry. Philadelphia (1908), 101. 



MANILA OOPAL. 199 

leave an oily stain on paper, the glycerides must practically be entirely 
decomposed. 

Furthermore, free glycerol is not soluble in cold linseed oil; hence 
that which is formed must be removed, otherwise its separation would be 
at least partly responsible for the cloudiness of undercooked varnishes. 

According to Sabin,^' "Each varnish maker has his own secret methods for 
treating the oil, which are probably all about alike. One of the most common 
methods is to heat the oil from 260° to 288° for a very short time, which seems 
to char certain impurities and coloring matter which settle out during the 
subsequent tanking or storing which it receives.** 

Gill,^' referring to bleached oil, says that "it is prepared by special process kept 
jealously guarded for the use of varnish makers. It may be prepared by heating 
linseed oil hotter than in the preparation of "boiled oil," to 260° to 300°, or 
by forcing oxygen through the oil." 

All fatty oils are decomposed above 250°, below this temperature they 
are little changed, above it the change is greater the higher the temper- 
ature. The products of the decomposition of neutral fatty oils, drying 
or nondrying, contain acrolein, a decomposition product of glycerol, 
hence free fatty acids must be formed. It is believed that one of the 
disadvantages of heating linseed oil at the temperature employed in 
varnish manufacture is the liberation of linoleic acid, which dries more 
slowly than linolein. It would seem that the corresponding amount of 
glycerol produced offers a more logical reason for the claims that oils 
containing free acids remain tacky. 

EXPfiRIMENTAL. 

One hundred grams of Faw Manila copal were boiled with 200 cubic centimeters 
of turpentine, in which this amount is only partially soluble, 100 cubic centimeters 
of raw linseed were then added through the reflux condenser and the heating 
continued with apparently.no further solution. 

The experiment was repeated in the same manner with the exception that 
the linseed oil contained 26 per cent by volume of oleic acid; upon continued 
boiling a complete solution of the resin took place. 

A quantity of the mixed fatty acids of linseed oil was prepared and added in 
varying proportions to raw linseed oil, depending upon the quantity of tinmelted 
resin it was desired to dissolve, and it was found that raw or boiled linseed oil, 
containing the free, mixed, fatty acids of linseed oil in the proportions of 10 
to 30 per cent calculated as oleic acid, formed homogeneous solutions with raw 
or fused Manila copal when the latter is added in the proportion of 10 to 30 
gallon varnishes and heated for a time at a maximum of 200°. When the tur- 
pentine was added before the oil, the boiling point of turpentine, 166° to 165°, was 
sufficiently high to eifect complete solution with the exception of such foreign 
matter as may be present in the resin. The subsequent addition of turpentine to 
the oil and res^n did not produce any cloudiness. 

'^Loc. cit,, 72. 

«*Oil Analysis. Philadelphia (1898), 86. 



200 RICHMOND. 

CHABACTKRISTICS OF VARNISH FILMS MADE IN THIS MANNER. 

The varnishes were filtered hot and allowed to stand in closed jars for 
three months. 

Well-planed strips of native hard wood were first given one coating of 
boiled oil and allowed to dry in a dust-free atmosphere. The surfaces 
were then thoroughly rubbed with ground pumice and the varnishes 
spread with- a brush. When the first coat was dry, it was in turn scraped 
and pumiced and a second coat applied. 

[Jiider the local conditions of heat and moisture, the varnish films 
which contained their ingredients in the proportion corresponding to 
approximately 8, 10, 12 and 15 gallon varnishes have now stood a year 
without showing any appreciable loss of luster. They were completely 
dry to touch and rubbing tests in twenty-four hours. 

Varnishes made with the proportion of oil corresponding to 20 and 30 
gallon varnishes did not dry completely in forty-eight hours, and wlien 
dry presented a dull appearance. All the varnishes made in this manner 
were exceptionally pale and did not tint or tone the color of either light 
or dark hard woods. 

Within the period of observation the films have not shown any tendency 
to check, as would be expected with a Manila copal varnish, short of oil. 

A few varnishes were made containing elemi resin in addition to 
Manila copal. The elemi ^^ resin used consisted of the residue remain- 
ing when Manila elemi is freed from terpenes by distillation under reduced 
pressure. 

It is a light-brown solid, with a brilliant fracture, soluble in the cold 
in practically all organic solvents including turpentine, and amounts 
to about 70 per cent of the elemi. 

A^arnishes made with equal parts by weight of elemi resin and fused 
Manila copal remained tacky for several days. A varnish containing the 
resin in the proportion of 1 of elemi to 2 of copal dried* to a hard, very 
lustrous film in one to two days. A laboratory filing case of dark, stained 
wood was given one coat of this varnish, and it has remained brilliant for 
over two years. A limited amount of this elemi resin appears to increase 
the elasticity and brilliancy of a Manila resin varnish without affecting 
its drying properties or the hardness of the film. 

In this connection attention is called to an oleoresinous varnish pre- 
pared in this laboratory from exclusively Philippine raw materials. 
Lumbang oil ^^ {Aleurites moluccana Willd.) was substituted for linseed* 
as a drying oil. It was incorporated with commercial Manila copal in 
the usual manner and the solution was diluted to proper spreading con- 

** Bacon. This Journal, Sec. A (1909) , 4, 93, 

"^ Candle-nut oil, Chinese "tiing" or wood oil; This Journal, Sec. A (1907), 
2, 439. 



MANILA COPAL. 201 

sistency with turpentine distilled from the resin of the Benguet pine ^^ 
{Pinus insulark Endl.). A well-prepared strip of red narra wood was 
given two coats of this varnish, and it has remained exposed »in the 
laboratory for a more than a year without any noticeable loss of dur- 
ability or luster. 

CONCLUSION. 

The changes which take' place during the cooking of varnish are 
largely changes in the oil rather than the resin, i. e., it is possible so to 
treat linseed oil, either by boiling or by adding to it linseed-oil acids 
previous to its addition to the fused resin, that it will form a clear, homo- 
geneous mixture with the latter which will remain so upon cooling,. 
without subsequent heating to temperatures- greater than 150° to 200*'. 

"'This Journal, Sec. A (1909), 4, 231. 



THE DESTRUCTIVE DISTILLATION OF MANILA COPAL. 



By Benjamin T. Brooks. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. I.) 



INTRODUCTION. 



The investigation of the distillation products of Manila copal was 
undertaken with the hope that some light would be thrown on the chem- 
ical changes accompanying the process of manufacture of the resin into 
varnish. It was expected that the distillation products themselves might 
furnish some clue as to the nature of the substances in the resin. 

The success of modern varnish manufacture depends upon the ex- 
perience and skill of the operator rather than on a scientific knowledge 
of the materials used. The process of manufacture is essentially the 
same as that described in books printed long before the development of 
modern chemistry. 

There is little accurate information available regarding the practice 
of modern varnish manufacturers. Although the resins used in the 
manufacture of resin oil varnishes are derived from many different 
species of trees and are of different ages when they come into the market, 
the process of manufacture is practically the same in all cases. 

In order to make the resins soluble in linseed oil they are heated, 
during which process they- loose considerably in weight. The older fossil 
resins are heated until they suffer a greater loss of weight than the softer, 
more recent ones, but the statements as to what per cent of the whole 
must be volatilized in each case vary within wide limits. 

Scheibler* subjected "copal from Manila" to dry distillation and noted two 
distinct stages in the process, the first being characterized by frothing and the 
presence of water in the distillate. 

Tschirch ^ distilled two samples of Manila copal. The properties of the harder 
sample described by him agree well with those of the material being studied in 
this laboratory. He states that the dry distillation of the hard sample yielded 
formic, acetic, and succinic acids among the products of decomposition. 

Friedburg" noticed that certain, copals yielded a distillate having an odor 

^Ann. d. Chem, (Liehig) (1860), 113, 338. 
^Arch. d, Pharm. (1902), 240, 202. 
'Journ. Am. Chem. 8oc. (1890), 22, 286. 
95193 3 * 203 



204 BROOKS. 

suggestive of limonene, but he was unable to isolate the latter. He does not 
state what variety of copal had this property. 

Wallach * subjected Kauri copal to destructive distillation and succeeded in 
isolating pinene and dipentene. 

L. Schmolling" examined the distillates from Kauri and Manila copals. The 
oils from the latter resin were refractioned, but no terpenes nor any of their 
characteristic derivatives were isolated. The aqueous portion of the distillate 
gave reactions for formic and acetic acids. 

The Manila copal used in these investigations was selected from the 

first and second grades furnished by the Manila Trading and Supply 

30° 
Company. The specifip gravity (tko) varied from 1.060 to 1.067. It 

has been pointed out by Tixier ^ that it is impossible to assign definite 
melting points to the copals."^ Powdered Manila copal softens suffi- 
ciently quickly to run together at about 100°. The freshly powdered resin 
possesses a pleasant^ aromatic odor which is probably due to limonene and 
pinene, as will be shown in the experimental part of this paper. 

The Koettstorfer number/ obtained by boiling with alcoholic caustic 
potash for one-half hour and using phenolphthalein as an indicator, 
averaged about 160, being a little higher for the small pieces and chips 
than for the large kimps. 

A study of the products given off by Manila copal during destructive 
distillation up to the point at which the frothing ceases and the melt 
is quite fluid has been made. It has been found that the decomposition 
takes place in two stages. The first is characterized by much frothing 
and the temperature of the mass rises slowly to about 330°, when it 
becomes fluid and appears to boil quietly. At this point the loss in 
weight is usually from 12 to 14 per cent of the original sample. The 
soft, white pieces sufl:*er a greater loss in weight than the very hard 
ones, due to the larger amount of volatile oils contained in them. The 
temperature of the gases entering the condenser usually does not exceed 
200° until the temperature of the melt reaches about 340°, at which 
point resin oil begins to distill over in large quantities. The temperature 
of the gases then quickly rises to about 330° and continues to rise slowly 
as higher boiling resin oils distill over. High-boiling resin oils are 
also carried over as a heavy mist with the gases which are rapidly 
evolved. 

*Afm. d. Chem. {Liehig) (1892), 271, 308. 

^Chern. Ztg. (1905), 29, 956. 

"^Joum, Soc. Chem. Ind. (1906), 25, 996. 

^Coffignier, Bull. 8oc. chim. Paris (1908), (4) 3, 453. 

•As Worstall has already shown, Manila and certain other copals are probably 
free from esters. The term Koettstorfer number implies no knowledge of the 
chemical structure of the resins and is here preferred to either saponification or 
acid number. 



DESTRUCTIVE DISTILLATION OF MANILA COPAL. ?0§ 

The products which are given off by the fused resin up to 330° are 
resin oil, pinene, p pinene, limonene, dipentene, camphene, water, formic 
acid, acetic acid, methyl alcohol, acetone, acetyl formaldehyde, formal- 
dehyde, furfuraldehyde, carbon dioxide, saturated and unsaturated hydro- 
carbons, and probably acrolein. 

THE TERPENES CONTAINED IN MANILA COPAL. 

The weight of oil obtained from 1,500 grams of small pieces was 94 
grams, of which about 1 cubic centimeter boiled below 150"^. About 
24 grams of the oil boiled between 150° and 178°, but the greater part 
of the fraction passed over between 170° and 178°. Owing to the small 
quantity of the lower boiling portion, a thorough study of it was not made. 

The presence of limonene in the fraction boiling between 170° and 
178° was shown by isolating limonene tetrabromide, a good yield being 

30° 

obtained. The oil showed an optical rotation of A-p-=58°.50. The 

low optical rotation is indicative of dipentene, but after several fractional 
crystallizations of the limonene tetrabromide the melting point, 104°, 
was unchanged. The crystals were dextrorotatory. Limonene nitrosyl 
chloride was prepared from this fraction and identified microscopically.® 
Crystals of pinene nitrosyl chloride were identified in the same manner 
in the lower boiling fraction. 

Limonene does not commonly occur in coniferous resins although it is found, 
accompanied by pinene, in pine needles. I limonene is obtained commercially from 
the needles of Ahies alba Miller. Recently Herty and Dickson" found I limonene 
to be the principal volatile oil in the oleoresin of Pinua serotina Michx. 

The presence of pinene in Manila copal was first noted by Mr. G. F. 
Eichmond,^^ who obtained it by distillation with steam. The volatile 
oil from certain specimens of Manila copal appears to consist almost 
entirely of pinene, while others show a high percentage of limonene. 
The presence of limonene could not be proved in the distillates from 
certain samples, and if present, it exists there in very small amount. 

As would be expected, individual pieces of the resin vary in the per- 
centage of volatile oils contained in them. It would probably be safe 
to assume that the soft, white resin, from* which about 11 per cent of 
terpenes was obtained, is of comparatively recent origin; while the 
hard, brown resin containing less volatile oil is much older. In general, 
very large lumps contain more volatile oil than small pieces. Large 
pieces which were quite hard on the surface and had a much weathered 
appearance were found to yield as high as 6 per cent of terpenes. The 

'Chace, Journ. Am. Chem. 8oc. (1908), 30, 1476. 
' ^^ Journ. Am. Chem. 8oc. (1908), 30, 872. 
'^This Journal, Sec. A (1910), 5, 186. 



206 



BROOKS. 



content of volatile oil therefore does not always indicate the real age 
of the resin. Eesin on the surface and interior of the same piece may 
be of quite different age with respect to the chemical changes which 
have taJten place during the ageing process. 

In order to investigate the terpene fractions more fully, several 
distillations were made on different specimens. The fractions boiling 
between 154° and 180° were as follows: 



Small, hard, light-colored pieces 

Large, hard, dark-colored lump with very friable surface 

Large, hard, light-colored lump 

Fresh, soft, white pieces 



Per cent. 

3.6 

6.0 

10.0 

n.2 



The volatile oil obtained from the soft, white material was shaken 
out with caustic soda, dried, and distilled over sodium. The following 
fractions were obtained : 



155°-1620. 


1620-168°. 


1680-170°. 


1700-180°. 


Residue. 


Grams. 
145 


Grams. 
16 


Grams. 
6 


Grams. 
12 


Grams. 
6 



The fraction from 155° to 162*=* consisted largely of pinene. It possessed the 

Q()o *^n° '^no 

following constants: Specific gravity ^=0.8487; N-— = 1.4672; A'^"- = 12.10°. 

o\) JL) D 

The presence of pinene was proved by the preparation of pinene nitrosyl chloride, 

which was identified microscopically. The nitrosyl chloride was then treated 

with benzylamine and the crystalline pinene nitrobenzylamine identified by its 

melting .point, 122°-123°. This fraction was redistilled twice over sodium and 

about 90 grams were obtained which boiled between 155°. 5 and 158° and showed 



the specific gravity —=0.8486; N^^= 
ou D 



:1.4655; A^ = 15°.50. 



It was always noted in the distillates from all the samples that a large portion 
of the terpene fractions . boiled from 160° to 1.65°. The fractions boiling from 
160° to 170° which were obtained from all the samples were united and frac- 
tioned. 

The fraction boiling from 160° to 161° was tested for camphene according to 
the method of Bertram and Wahlbaum." On slowly subliming the mixture of 
terpineol and isoborneol which this method yielded, crystals of isoborneol were 
obtained. Nothing approaching a quantitative separation of the substances could 
be made. The crystals which were deposited were very characteristic, forming 
six-pointed star-shaped crystals of the hexagonal system. They appeared to be 
identical with those obtained by subliming a synthetic mixture of isoborneol and 
terpineol. The crude isoborneol was recrystallized once from ligroin, but the 
quantity obtained was not sufficient to permit of further purification. The crystals 
melted at 198° to 204.° 



^Joum. f. prakt. Chem. (1894), (2) 49, 15. 



DESTRUCTIVE DISTILLATION OF MANILA COPAL. 207 

The fraction boiling between 164° and 166° was tested for /3 pinene by oxidizing 
with alkaline permanganate in the cold.*' From a sample weighing 30 grams only 
a very small quantity, about 0.5 gram, of nopinic acid was isolated, which after, 
recrystallizing from ligroin melted at 124° to 125.° About 16 grams of the oil 
was unacted upon and was recovered. This was again treated with alkaline 
permanganate and a small quantity of nopinic acid again obtained. 

The fraction boiling from 165° to 170° was tested for phellandrenfe, but with 
negative result. 

The fraction boiling from 170° to 180° was also examined for aylvestrene by 
testing for its color reaction with acetic anhydride and a few drops of sulphuric 
acid. The result was negative. An attempt to isolate sylvestrene hydrochloride 
from the hydrochlorides of the hydrocarbons in this fraction also gave negative 
results. 

No limonene tetrabromide could be isolated from the fractions of any 
of the distillates boiling from 170° to 180° and taken from samples 
yielding 10 per cent or more of terpenes.^* Limonene was found only 
in the distillates from samples yielding from 2 to 3 per cent of volatile 
oil. I hardly feel warranted in advancing any explanation of these facts. 
Dr. F. W. Foxworthy/^ of the section of botany, biological laboratory, 
Bureau of Science, is convinced that almaciga resin is produced by one 
species of tree only. Whether certain individual trees of this species 
produce limonene and others do not, as was found' to be the case with 
Manila elemi,^^ has not been determined. Wallach ^' believes that limo- 
nene is formed in the plant from pinene in the presence of acids. This 
question may be taken up more fully at a later date. 

THE GASES GIVEN OFF BY MANILA COPAL DURING DISTILLATION. 

The gas given off by the copal during fusion was collected and analyzed. 
The carbon dioxide in the gas was determined quantitatively at intervals 
during the distillation. Below 330° the percentage of carbon dioxide 
remained practically constant at 92.6' per cent, as was demonstrated by 
repeated analyses. When the temperature had reached 330° the change 
in the nature of the decomposition was shown by the difference in the 
composition of the escaping gas. Beginning at this point, gas analyses 
made at intervals of ten minutes showed the following results: 

^'Baej^er and Villiger. Ber. d. deutschen chem. Oes. (1896), 29, 22. Schimmel 
& Co., Rep. (1908), April, 103. 

** The only terpene positively identified in these fractions was dipentene. The 
fractions were brominated as for limonene and the solutions placed in an ice box 
for four days. Only very small yields of crystals were obtained, which, after 
recrystallizing from alcohol and ether, melted at 120*^. 

^^This Journal, Sec. A (1910), 5, 

^•Clover, This Journal, Sec. A (1907), 2, 1; Bacon, This Journal, Sec. A (1909), 
4, 93. 

''Ann. d. Chem. {Liehig) (1888), 246, 235. 



208 BROOKS. 

Table I. — Analyses of the gases evolved after the first stage of the distillation. 



Number of determination. 


Carbon 
dioxide. 


Unsaturat- 
ed hydro- 
carbons. 


Saturated 
hydro-car- 
bons and 
CO. 


1, 


92.6 
75.4 
38.8 
28.8 

28.8 






2 


6.4 
15.0 
14.5 
14.6 


18.2 
46.2 
56.6 
56.6 


.8 


4 


5 __ 





The total quantity of carbon dioxide given off during the first stage 
was determined in another experiment in which this gas was absorbed by 
a solution of caustic soda. An aliquot portion of the alkaline solution 
was acidified in the form of apparatus described by Hillebrand ^^ for the 
determination of carbon dioxide in carbonate rocks, the gas being absorbed 
in caustic potash and weighed. The amount found corresponded to a 
total of 48 grams from 1,500 grams of resin, or 3.2 per cent. 

This amount of carbon dioxide is probably given off by the decomposition of a 
carboxyl group. The resin oil which distills over when the temperature is raised 
consists largely of neutral oils. One sample of resin oil contained 63 per cent 
of neutral substance. As has been shown by G. F. Richmond ^® the original resin- 
ous substance consists almost entirely of compounds of an acid character. The 
ease with which these resin acids lose carbon dioxide is similar to the behavior 
of abietic acid, Avhich can not be distilled even in vacuo without the loss of a 
large percentage of that gas. 

The gases given off during the melting of resins used in varnish making 
are commonly described as having an irritating odor, but are not gener- 
ally considered as being dangerous. 

Bachem^° recently recorded a case, of fatal poisoning by the gases from a 
melting kettle in a varnish factory. He analyzed the gases given off by colophony 
and found carbon monoxide in amounts varying from 10.6 to 39.0 per cent. 

The amount of carbon monoxide given off by Manila copal was quan- 
titatively determined by absorbing the gas in cuprous chloride, in the 
usual manner, care being taken that the unsaturated hydrocarbons were 
previously removed by strongly fuming sulphuric acid. Two samples, 
taken at different stages, showed 11.4 and 15.5 of carbon monoxide. The 
presence of this gas was also shown by killing two- white rats with small 
amounts and examining the absorption spectra of samples of their blood 
before and after the addition of a few drops of ammonium sulphide. 
Considering the rather large amounts of carbon monoxide found in the 
gases, it is surprising that their toxicity has not been more often noted 
by varnish manufacturers, or their workmen. 

«?7. S. GeoL 8urv. (1907), Bull. No. 305, 151. 

'"This Journal, Sec. A. (1910), 5, 187. 

"^ Arch. f. exp. Path. u. Pharm. tl907), 57, 222. 



DESTRUCTIVE DISTILLATION OF MANILA COPAL. 209 

THE LIQUID PRODUCTS GIVEN OFF DURING THE DISTILLATION OF MANILA 

COPAL. 

Water is evolved continually and smoothly below 330° and appears 
to result from chemical decomposition. Powdered samples dried over 
phosphorus pentoxide or in vacuo over sulphuric acid readily yield water 
on heating. The aqueous solution obtained ^ from the distillation of 
1,500 grams of resin weighed 46 grams. The substances soluble in water 
which were found dissolved in the aqueous layer weighed about 9 grams, 
leaving 37 grams for water, or 2.4 per cent of the weight of resin taken. 

The volatile acids detected in the aqueous layer were formic and acetic acids. 
The aqueous solution was drawn off and the oil washed three times with water, 
the washings being added to the original aqueous solution. The solution was 
then diluted to 500 cubic centimeters. Formic acid was detected by the ferric 
chloride test ^^ and determined quantitatively in an aliquot part of the solution 
by the mercuric chloride method.'''* From the amount found the total was cal- 
culated to be 4.6 grams. The remainder of the acidity due to volatile acids, when 
calculated as acetic acid, was equivalent to 3.1 grams. 

Formaldehyde was shown to be present by the well-known reaction with milk 
and sulphuric acid containing a trace of ferric chloride. A previous distillation 
had shown that all the formaldehyde was condensed with the water, so that a 
quantitative determination of it was made in an aliquot part of the solution by 
the ammonia method as described by Williams.^-^ The amount found corresponded 
to a total of 1.13 grams. 

The aqueous solution remaining after the acids and formaldehyde had been 
determined amounted to 250 cubic centimeters. This was treated in the cold 
with 1 cubic centimeter of phenylhydrazine dissolved in water and the mixture 
allowed to stand in an ice chest over night. The light, crystalline precipitate 
which separated was filtered, washed with a little cold alcohol, and dried over 
sulphuric acid. The crude crystals weighed 0.22 gram and melted at 135°. 
After recrystallization once from hot alcohol and once from benzene, the substance 
melted at 143° to 144°. A combustion was also made on the body; 0.1515 gram 
substance gave 0.3960 gram CO2 and 0.0855 gram H2O. 





(Calculated for 






C]5Hi6N4. 


Found 




Per cent. 


per cent. 


H 


6.34 


6.27 


C 


71.42 


71.28 



The crystalline substance is therefore the osazone of acetyl formaldehyde recently 
described by Dennis.^* 

Furfuraldehyde was detected in the aqueous solution by its reaction with 
metaxylidene acetate.^^ From the work of Aschan^ on the distillation of wood 
turpentine, diacetyl and some of its homologues were expected, but were not found. 

Small traces of other aldehydes are probably present, as an oil is formed in 

2* Smith, Journ.. Am. Chem. Soc, (1907), 29, 1236. 

"Merl. Ztschr. f. Untersuch. d, Nahrungs-u, OenussmitteL (1908), 16, 385. 

^' Journ. Am. Chem. 80c. (1905), 27, 597. 

2* Aw. Chem. Journ. (1907), 38, 583. 

2°Schiff. Ann. d. Chem. {Liehig) (1887), 239, 380. 

'''Ztschr. f. ang. Chem-. (1907), 20, 1811. 



210 



BROOKS. 



addition to the crystalline substance, when the cold aqueous solution is treated 
with phenylhydrazine. The irritating odor of the original distillate is very sug- 
gestive of acrolein, but its presence was not proved. 

The presence of a very small amount of methyl alcohol was shown by neutraliz- 
ing the aqueous, acid solution with caustic soda and distilling over about one- 
third of the solution. The distillate was saturated with potassium carbonate and 
redistilled. The first small fractions Were then tested for methyl alcohol by 
oxidizing to formaldehyde with hot copper oxide and then testing for the aldehyde. 

The presence of acetone among the distillation products was shown in another 
experiment in which the distillation products were treated with an excess of 
dilute caustic-soda solution and distilled. The first few cubic centimeters were 
tested for acetone by the method of Gunning " in which test iodoform is formed 
on the addition of a solution of iodine in ammonium iodide and a few drops of 
dilute ammonium hydroxide. This was confirmed by the test depending upon the 
ability of acetone to dissolve mercuric oxide in an alkaline solution.-* 

The products given off by Manila copal up to the point at which the 
melt has a temperature of about 330° may thus be summarized. 

Table II. — Suhsta/nces given off hy Manila copal during the first stage of the 
decomposition, up to SSO"". 



Substance. 



Carbon dioxide 

Water .— 

Formic acidi 
Acetic acidj 
Formaldehyde, acetyl formaldehyde, furfuraldehyde, 

methyl alcohol, and acetone, approximately. 

Gaseous hydrocarbons 

Pinene, limonene, dipentene, |8pinene,and camphene, 

variable. 
Resin oil, variable, usually from 



2.4 
0.5 
0.2 

0.2 
1.5-11.2 



DISCUSSION or RESULTS. 

The quantitative results recorded in this portion of the paper should 
be considered as those obtained from commercial Manila copal of average 
age and of the first and second grades. Any discussion of the properties 
of copals would have to take into account the wide differences in the age 
of the specimens examined and the consequent variations resulting there- 
from. The differences in the yields of terpenes obtained have already 
been pointed out. Since the resin absorbs oxygen quite rapidly and 
increases in acid value it is to be expected that the amounts of carbon 
dioxide, water, and other substances will vary according to the age of the 
resin which is distilled. 



"Ztschr. /. anal. Chem. (1885), 24, 147. 

^Gunning and Reynolds, Ztschr. f. anal. Chem. (1885), 24, 148. 



DESTRUCTIVE DISTILLATION OP MANILA COPAL. 211 

It is at once apparent that the carbon dioxide, oil, and water are the 
only products given oif which can play an important p^rt in the chemical 
changes involved during the first stage of the decomposition, 

Xhe products formed are similar to those obtained by the incomplete combustion 
of hydrocarbons.^' 

Formic and acetic acids, acetone, acetaldehyde, furfuraldehyde, ethylene, pro- 
pylene, carbon dioxide, and methane have repeatedly been found among the 
products of decomposition by heat of the more complex carbohydrates and other 
organic compounds." The formation of these substances during the distillation of 
Manila copal can not therefore be considered as throwing much light on the 
constitution of the original material. Their production suggests that possibly 
a carbohydrate group may exist in some of the original resinous substance. This 
is made somewhat more probable by the observation of Strasburger ^^ that in the 
conifers the starch content of the wood parenchyma cells is displaced by resin. 
In other words, the resin appears to be formed at the expense of starch, The 
same author says: "Since the conifers sacrifice considerable quantities of reserve 
material in order to produce resin, it is clear that the starch is transported to 
the place at which the resin formation shall take place. 

The darkening of the lesin on melting is not caused by oxidation, but 
appears to be due largely to the carbonization of some of the constituents. 
When the distillation is performed in glass retorts, the air is quickly 
displaced by carbon dioxide, steam, and gaseous hydrocarbons. Small, 
black pieces of charred substance can always be seen in the clear melt. 
Samples heated in sealed tubes to 300° darken to the same degree as 
when heated in a retort. A sample prepared by .the evaporation of a 
clear, alcoholic solution, from which the rather insoluble flocculent matter 
had been removed, was heated in a sealed tube. The darkening was 
less in this case than when the original resin was employed. 

The distillation was not studied beyond the first stage except to identify 
ethylene and propylene among the gaseous hydrocarbons. 

This was done by passing the gases through bromine water. The oil thus 
obtained was distilled and a fraction weighing 5 grams was obtained which 
boiled from 128° to 132°. In one experiment all of the unsaturated hydrocarbons 
excepting ethylene were removed by passing the gas through concentrated sul- 
phuric acid. The ethylene dibromide obtained in this case boiled between 128° 
and 134°. Propylene was identified in the fraction boiling from 135° to 145° 
by heating the oil with ten times its weight of water in a sealed tube at 180° 
for eight hours. The aqueous solution thus obtained yielded iodoform on treat- 
ment with sodium carbonate and iodine, indicating acetone, and an oily pre- 
cipitate separated on the addition of phenylhydrazine, which when heated with 
zinc chloride gave off the odor of skatol, indicating propionic aldehyde.*^ 

2»Stepski, Monatsh. /. Chem. (1902), 23, 773. 

«> Fisher, E. Ber. d. deutschen chem. Ges. (1889), 22, 105; Arm. d, Chem. 
{Liehig) (1890), 258, 230; Pinner, Ber. d. deutschen chem. Oes. (1883), 16, 
1730. 

"^ Histologische Beitrdge, Jena (1901), 3, 4. 

»2 Fisher, E. Ber. d. deutschen chem. Ges. (1889), 22, 353. 



212 BROOKS. 

Further investigations of the products given off during the second 
stage of the decomposition were not undertaken because of the impos- 
sibility of drawing any conclusions from them as to the nature of the 
original substances. The complete distillation of Manila copal is more- 
over not a conimercial practice. 

The fact that high-boiling resin oils are carried over by the gases 
shows that a certain amount of resin oil must be left in the melt. The 
presence of resin oil in the melt may in part account for the solubility 
of the melted resin in linseed oil. 

No chemical examination of the copal oil has been made. The fractions boiling 
from 190° to 200° probably contain oxidation products of the terpenes contained 
in the resin. These fractions possess a fine, camphor-like odor and have specific 
gravities ranging from 0.880 to about 0.910. Guedras ^^ has suggested that copals 
may contain oxidation products of the terpenes. 

Numerous patents have been taken out on processes which have for 
their object the minimizing of the loss in weight caused by the heating 
of the resins. 

In some processes the resin oil is condensed and returned to the melting kettle."* 
According to other methods the resins are heated under pressure."' The resin 
oil thus retained in the melt undoubtedly assists in the solution of the resin 
and less heating is probably required than when the oil is lost. 

THE EFFECT OF HEATING MANILA COPAL ON ITS SOLUBILITY IN OILS. 

The formation of copal oil during destructive distillation at tem- 
peratures which did not exceed the common working temperature of 
varnish manufacture suggested that perhaps the presence of this oil 
in the melted resin plays an important part in the solution of fused 
copal in linseed oil. 

I have found that copal oil is formed at the common working 
temperature of varnish manufacture. It assists in the solution of the 
resin in linseed oil, and samples of melted resin containing a large per 
cent of copal oil can be dissolved in greater quantities of linseed oil, 
without separating on cooling, than those having a smaller per cent. 
If the copal oil is removed from a sample of fused resin by distilling 
under reduced pressure, the solubility of the residue in linseed oil is 
much diminished. The substance thrown out of a solution in ligroin 
or benzene by diluting with ligroin contains no resin oil and is extremely 
insoluble in linseed oil. It is converted into a dark oil if further 
decomposed by strong heat, or to pale oil varnish at much lower tempera- 
tures by the use of partially saponified linseed oil.^^ 

^Compt. rend, Acad. set. (1902), 135, 797. 
»*Tixier, Journ. Soc. Chem. Ind, (1906), 25, 996. 

"Hempel, Journ. Soc. Chem. Ind. (1901), 20, 818; Smith, Ihid. (1901), 20, 
1076; Lippert, Ihid. (1901), 20, 1122. 

"Richmond. This Journal, Sec. A (1910), 5, 199. 



DESTRUCTIVE DISTILLATION OF MANILA COPAL. * 213 

EXPERIMENTAL. 

The following experiment shows that copal oil is, as a matter of fact, 
formed by the decomposition of Manila copal at the common working 
temperature, about 300°. 

About 250 grams of resin were heated to 310** and a reduced pressure of 120 
millimeters applied. Even with this slight vacuum 26 grams of a viscous, light 
yellow oil were obtained. After this amount of oil had been distilled, the heating 
was continued at ordinary pressure for four hours at 300° to 310". When 
reduced pressure was applied, as before, 28 grams of oil were obtained below a 
temperature of 290°. 

A sample of Manila copal was heated until it had lost about 20 per cent of 
its weight. A portion of the fused residue was reheated to 200** and an equal 
quantity of hot, boiled linseed oil added to it. When a drop of the mixture was let 
fall on a piece of glass and allowed to cool it remstined clear, but after adding more 
linseed oil, amounting to about 25 per cent of the original quantity of oil, the 
drop then taken became cloudy on cooling, owing to partial separation of the 
resin. On adding still more hot boiled linseed oil the resin was precipitated in a 
bulky sticky mass. ' 

It will be seen from the above experiment, which was duplic!ited on 
many different melted samples, that the behavior of linseed oil as a 
solvent for heated copals is the same as that of ligroin, turpentine, and 
benzene. Varnish-like solutions can also be made of melted resin in 
coconut oil and heavy mineral oil which boils above 300°. The resulting 
solutions resemble yarnish in their transparency, but of course do not dry. 

The resin was melted exactly as in the preparation of a linseed-oil varnish 
and heated at 300° until it had lost about 25 per cent of its weight. The hot 
mineral or coconut oil was then added in the proportions required for a lO-gallon 
varnish and the heating was continued at about 290° for one-half hour. The 
drops then tested remained quite clear on cooling. A portion of each solution 
was thinned with an amount of turpentine equal to the quantity of oil taken. 
The addition of more mineral or coconut oil to the undiluted portions caused 
partial precipitation of the resin. 

It is naturally to be expected that the loss of carbon dioxide and 
w^ater which occurs on fusing the resin will affect its solubility. The 
melted resin is somewhat richer in hydrocarbons and poorer in acids 
than the original material. The Koettstorfer numbers of melted copals 
are always lower than those of the natural resins. The decrease in the 
Koettstorfer number depends upon the duration and temperature of, 
and probably, the pressure exerted upon the fusion. However, it does 
not seem to me at all probable that the loss of relatively little carbon 
dioxide, water, and small quantities of other substances ^^ would, in itself, 
affect the solubility of the copal to any considerable extent. . 

^ See page 210. 



214 ' BROOKS. 

THE FORMATION OF COPAL OIL AT TEMPERATURES ABOVE 320°. 

The formation of copal oil is much more rapid at a. temperature of 
320° or above, than it is at 300°. Between 320° and 360° large quantities 
of oil distill from the melt. 

Many resins may be melted with very large proportions of oil from 
colophony or copal and yet solidify to hard masses. Eesidues from 
vacuum distillations, which contain very little or no copal oil, are quite 
brittle. The 'addition of copal oil makes the mass tougher. 

A sample of Manila copal was melted with an equal weight of copal oil. On 
cooling to 30° the mass became sufficiently hard to break with a conchoidal 
fracture when struck a sharp blow. Synthetic mixtures cont&,ining 25 per cent 
of copal oil are quite hard at ordinary temperatures. 

Copal oil is miscible in aJl proportions with absolute ethyl alcohol, 
amyl alcohol, acetic acid, acetone, ether, chloroform, ligroin^ benzene, 
tu'rpentine, and linseed oil. When a clear, concentrated solution of 
heated copal in one of these solvents is partially precipitated by diluting 
with the same solvents, the copal oil is left in solution together with 
some resinous substances, while the precipitated portion is free from 
copal oil. The effect of removing the copal oil in this manner is shown 
by the following experiments. 

A sample of Manila copal weighing 500 grams was melted in a covered iron 
vessel and the temperature maintained at about 320° to 325° for one-half hour. 
The loss in weight after this treatment was 25.0 per cent."' A portion of this 
melted sample weighing 50 grams was remelted, the temperature not rising above 
220°, and 71 grams of hot boiled linseed oil, which had previously been tested for 
purity, was stirred in. These proportions correspond to the composition of a 
10-gallon varnish, the calculation being based on the weight of the unheated 
resin. When all of the oil had been added, a drop test showed no cloudiness and 
a small test tube sample remained clear on cooling. When the solution had 
cooled to 170°, 25 cubic centimeters of turpentine were added. A sample of 
this mixture also remained clear after cooling. 

This experiment was performed for the sake of comparison with the 
following : 

A portion of the same melted sample weighing 100 grams was completely 
dissolved in 100 cubic centimeters of benzene. The resin was then partially pre- 
cipitated by adding 800 cubic centimeters of ligroin. The solvent was decanted 
from the precipitate and evaporated on the steam bath and finally over a free 
flame until the temperature of the mass had risen to 18P° and all of the solvent 
had been expelled. The residue weighed 65 grams and evidently contained a 

•^A large part of this loss in weight represents volatile oils. Obviously, the 
loss in weight accompanying a given amount of decomposition will depend more 
or less on the form of the containing vessel and the manner of heating. I venture 
the opinion that the old drop test gives a more accurate indication of the point 
beyond which further heating is unnecessary than any arbitrary rule such as 
heating until a certain per cent of weight is lost. 



DESTRUCTIVE DISTILLATION OP MANILA COPAL. 215 

considerable quantity of high-melting, resinous substance in addition to the copal 
oil. It melted at approximately 75° to 80''. This residue was melted and mixed 
with linseed oil, and finally thinned with turpentine, in the same proportions 
Used in the above experiment. Samples taken before and after the addition of 
turpentine remained , clear after cooling. 

It was found that the precipitated portion gave up the solvent by spontaneous 
evaporation. In a few hours it could be crumbled between the fingers and after 
twenty-four hours was completely converted into a dark brown powder weighing 
35 grams. The substance melted at 185° to 190°. It was melted and heated 
to 230° and the same proportions of linseed oil added as in the first case. Pre- 
cipitation occurred even at this temperature and test-tube samples, taken before 
and after the addition of turpentine, gave such an abundant fiocculent precipitate 
on cooling that the whole appeared as if gelatinized, and the tubes could be held 
for several minutes in an inverted position without the contents flowing out. 

Since the precipitation method gives only a partial separation of oil 
and resin, an attempt was made, to remove the copal oil by distilling 
under reduced pressure. 

A sample of melted resin, the same as was used in the above experiments, was 
melted and a reduced pressure of 20 millimeters applied. The sample, weighing 
150 grams, was heated gradually to 260°. The oily distillate collected up to this 
point weighed 27 grams, or 18 per cent of the amount taken. This was deemed 
sufficient to determine whether or not the solubility of melted copal was markedly 
affected by the oil. The distillate was a light yellow, viscous oil having a greenish 
fluorescence. The portion coming over at about 260° and 20 millimeters pressure 
had very much the consistency of soft vaseline. The residue left in the flask 
was extremely brittle and friable. 

A portion of the residue weighing 50 grams was melted and heated to 230°. 
Boiled linseed oil, heated to a temperature of about 200°, was then slowly added' 
until drop tests showed a marked and unmistakable cloudiness. A small test-tube 
sample taken at this point also showed an abundant precipitate on cooling. The 
copal oil which had been removed by distilling under diminished pressure, was 
then added in the proportions corresponding to the composition of the sample 
before distillation. The hot mixture was stirred for several minutes, and a drop 
then tested showed no cloudiness on cooling and small test-tube samples, taken 
before and after thinning with 25 cubic centimeters of turpentine, remained per- 
fectly clear after cooling. 

The amount of linseed oil added in the above experiment up to the point at 
which unmistakable precipitation occurred on cooling was 130 grams, which is 
about equivalent to the composition of a 21 -gallon varnish, the calculation being 
made on the weight of the sample before heating. 

For comparison, an experiment was made on a portion of the original 
melted sample used in the above experiments with the object of determin- 
ing how much linseed oil could be added to it before partial precipitation 
would take place. 

The resin was heated to about 230°, as in the previous experiments, and hot 
linseed oil slowly added until drop tests showed unmistakable cloudiness. The 
amount of linseed oil used up to this point was found to correspond very closely 
to the amount required for a 30-gallon varnish. 



216 



BROOKS. 



THE BEHAVIOR OF MANILA COPAL IN SOME OF THE COMMON SOLV ENTS. 

The solubilities in several of the more common solvents of Manila 
copal, as it comes into the market, have been determined. Those of 
four commercial grades of the resin were found to be practically the same 
in each given solvent. A sample which had been heated to 300° was 
taken. AVhen large proportions of solvents were used, the solubilities 
in ligroin, benzol, and turpentine appeared to be about the same as for 
unheated copal and in alcohol this factor was markedly less. 

The method was as follows: 

One hundred cubic centimeters of the solvent were taken to 10 grams of the 
powdered resin and the whole boiled under a reflux condenser for one-half hour. 
About 30 cubic centimeters of the clear, hot solution were then decanted into a 
tared evaporating dish and weighed. This was then evaporated in an atmosphere 
of carbon dioxide. From the weight of resin found in this way the per cent which 
had gone into solution was calculated. In the case of ethyl alcohol, amyl alcohol, 
and ether, all of the resin appeared to go into solution with the exception qf a 
little flocculent substance which remained undissolved even in the presence of 
a large excess of the solvent. W^ith the exception of this flocculent matter, much 
greater quantities of resin can be dissolved in 100 cubic centimeters of ether, 
ethyl and amyl alcohols than the amounts taken in the experiments described 
above. 

Table I. — Per cent of substance dissolved from 10 grams of resin hy 100 cubic 

centimeters of solvent. 



Solvent. 



Ethyl alcohol 
Amyl alcohol 

Ether 

Ligroin 

Benzene 

Turpentine — 



Per cent. 



Temperature 

used to expel 

solvent. 



130 

150-155 

120 

130 

135-140 

170-175 



The above figures are only approximate and represent averages, since 
the solubilities of different pieces of the same grade vary slightly. The 
terpenes, in some samples amounting to 10 per cent, will evidently be 
expelled with the turpentine and lead to considerable error when this 
solvent is used. 

Because gf the lack of any simple method for separating resin and 
linseed oil, no quantitative determinations of the solubility of the copal 
in linseed oil were made. It appears to be a substance of about the same 
solvent power as ligroin, turpentine, and benzene. 

As has been pointed out by Heupel,^* all attempts to dissolve copals without 
previous fusion have been unsatisfactory, at least commercially. The solubility 



'^Journ, Soc. Chem. Ind. (1901), 20, 818. 



DESTRUCTIVE DISTILLATION OF MANILA COPAL. 



217 



of fused Madagascar copal was investigated by CoflSgner,*® who noted a decided 
decrease in its solubility in alcohol and increased solubilities in ether, benzene, and 
chloroform. 

A sample of Manila copal was heated slowly to 300° and kept at that 
temperature until it had lost 20 per cent of its weight. The solubility 
of the melted resin was determined as before. 

Table II. — Per cent of substance dissolved from 10 grams of melted copal by 
100 cnbic centimeters of solvent. 



Solvent. 



Ethyl alcohol 

Benzene 

Turpentine __. 



Per cent 
of copal 
dissolved. 



72 
49 
45 



The solubilities of this sample of melted copal in benzene and tur- 
pentine^ in the above proportions, are about the same as those of the 
unheated resin. However, these factors for fused resin depend very 
much on the temperature and the duration of the heating. So-called 
superfused copal is quite soluble in benzene and turpentine. 

AVhen small proportions of solvent to resin are employed, the solubil- 
ities are strikingly different from those given above. Samples of Manila 
copal which had been heated to about 300° to 320° were completely 
dissolved in small quantities, one to two parts, of ligroin, benzene, or 
turpentine, whereas unheated copal could not be so dissolved. The 
amount of unheated resin dissolved by these solvents is, roughly, in 
direct proportion to the amount of solvent used, until all but the more 
insoluble, flocculent matter mentioned above has gone into solution. 
However, the amount of fused resin dissolved is not even roughly pro- 
portional to the amount of solvent used. In fact, when a perfectly 
clear^ homogeneous solution of fused resin in one to two parts of tur- 
pentine, ligroin, or benzene is diluted with the same solvent, the resin 
is partially precipitated. Copal which has been fused for several hours 
at about 300° will form clear, homogeneous solutions with larger pro- 
portions of these solvents than resin which has been so treated for only a 
short time. 

One specimen of copal was heated for eight hours at a temperature 
of about 320°. The resulting mass was rather soft at 30° and dissolved 
completely in five times its weight of benzene and almost no precipitate 
was obtained on diluting. This superfused resin apparently consisted 
largely of resin oil together with a little resinous matter and charred 
substance. It had about the same consistency as a synthetic mixture 
containing 50 per cent resin and 50 per cent resin oil. 



^""Bull. Soc. chim. Paris (1906), 35, 762. 



THE OXIDATION OF MANILA COPAL BY THE AIR. 



By Benjamin T. Brooks. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. /.) 



Since it is generally considered that the older copal resins are the 
more valuable for varnish manufacture, it appeared desirable to investi- 
gate the changes taking place in Manila copal during its oxidation in 
the air. 

Little is known of the chemical structure of the compounds present 
in fresh and old resins of the so-called copal type. A brief review of 
some of the properties of certain important constituents of the common 
resins will be useful in understanding the behavior of Manila copal. 

A very common constituent of the resins of the Coniferce is abietic acid. 
According to Frankforter * this acid, or one of its isomeric foums, is present in 
nearly all natural resins. It is easily obtained in a crystalline form and has 
been more thoroughly studied than any of the resin acids. 

One of the difficulties met with in the study of this and similar substances is 
the ready absorption of oxygen from the air, which led to the discussion as to 
whether its formula was CieHasOj or C20H30O2.* According to Fahrion,-^ the absorp- 
tion of oxygen is accompanied by the formation of peroxides. 

A large number of acids related to abietic acid, and in some cases 
probably identical with it, have been found in many different resins. 
It has not been isolated from Manila copal, but the behavior of this resin 
on oxidation by the air is entirely similar to that of colophony or abietic 
acid. The absorption of oxygen from the air appears to be a property 
common to all complex resin acids, so far as they have been studied. 

Klason and Kohler * studied the resin acids secreted by the Norway spruce 
and, prepared them so as to avoid heating. The acids isolated by them absorbed 
oxygen from the air with great rapidity, forming oxyacids which were quite 
insoluble in turpentine and were particularly adapted to forming a protective 
coating when the oleoresin exudes from the tree. The authors claim that it is 
not necessary for oxidation or evaporation to take place in order that the secre- 

'Journ. Am. Chem. Soc. (1909), 31, 563. 
2 Levy, Ztschr, f. ang. Chem. (1906), 18, 1739. 
^Ztschr. f. ang. Chem. (1901), 14, 1197; (1907), 20, 356. 
*Journ. f. prakt. Chem. (1906), N. F. 181, 337. 
95193 4 219 



220 BROOKS. 

tion harden,'' but there is no doubt that under natural conditions both oxidation 
and evaporation accompany the hardening process, beginning as soon as the 
fluid secretion is exposed to the air. 

The terpenes present in resins of this type also readily undergo oxidation by 
the airf and their tendency to resinify is much affected by the presence of other 
substances.* The changes during autoxidation will be discussed in another part 
of this paper. 

The effect of atmospheric oxidation on the chemical constants of freshly pow- 
dered samples of the more common varieties of copal was investigated by Wor- 
stall,^ who noted a marked increase in their acid numbers and a decrease in their 
iodine absorption. The greatest change was noted in the case of Kauri, while 
specimens of hard fossil Zanzibar resin did not change appreciably in four months. 
The greatest increase observed by him in the acid number of Manila copal over 
a period of four months was eleven units. 

I have found that the atmospheric oxidation of Manila copal is 
accompanied by the formation of organic peroxides^ an increase in the 
Koettstorfer number and evolution of small quantities of carbon dioxide, 
formaldehyde, formic acid, and hydrogen peroxide. The oxidation is 
accelerated by sunlight. Under certain conditions the resin gives off 
vapors which have the property of affecting a photographic plate. 

THE EVOLUTION OF CARBOX DIOXIDE FROM MANILA COPAL. 

In a study of the distillation products ^ of this resin, it was noticed 
that it was impossible to melt the resin without large quantities of 
carbon dioxide being given off. At temperatures between 250° and 
330° this gas is evolved in large quantities and is apparently derived 
from the decomposition of a carboxyl group. Easterfield and Bagley •' 
found that pure abietic acid gives off large quantities of carbon dioxide 
when distilled in vacuo. However, small quantities of carbon dioxide 
are given off by Manila copal during atmospheric oxidation even at 
ordinary temperatures. 

Schwalbe *® noted that colophony gave off carbon dioxide at about 120° and 
suggested that the Koettstorfer number might be affected by this cause. He 
regarded the gas as being derived from the decomposition of abietic acid. 

^ I have recently had the opportunity of making a similar observation in the 
case of the oleoresin secreted by the pines, Pinus insularis Endl., in the Province 
of Benguet in northern Luzon. Samples of the clear, honey-yellow exudate were 
poured into small vials and immediately sealed. In about one hour the contents 
of the vials had solidified to a mass having the consistency of stiff lard. Stirring 
the clear liquid greatly accelerated this hardening process which was due to the 
crystallization of the abietic acid, as shown by a subsequent examination of the 
substance. 

"Bacon, This Journal, i^ec. A (1909), 4, 93. 

^Journ. Am. Chem. 8oc. (1903), 25, 860. 

*Thi8 Journal, Sec. A (1910), 5, 

'Journ. Chem. 8oc. Lond. (1904), 85/ 1238. 

^""Ztschr. f. ang. Chem. (1906), 18, 1852. 



OXIDATION OF MANILA COPAL BY THE AIR. 221 

Herty and Dickson " have recently confirmed Schwalbe's experiment on speci- 
mens which had been exposed to the air, but find that resin which has not been 
long in contact with the air or oxygen can be heated indefinitely at a temperature 
of 140° without losing carbon dioxide, provided oxygen and moisture are excluded 
from the flask in which the resin is heated. Turpentine which had been exposed 
to the air gave off carbon dioxide when warmed. The authors state as follows : 
"No question of the splitting off of a carboxyl -group could arise here.** They 
also prepared specimens of the resin acids from the oleoresin of Pinus heterophylla 
(Ell.) Sudworth, in such a manner as to avoid heating. However, they make 
no mention of having made any effort to exclude oxygen. The salts of the resin 
acids were dissolved in water, the solutions acidified and the precipitated acids 
washed and dried. I have found that, when the resin acids from Manila copal 
or colophony are prepared in this manner, the fine powders obtained rapidly take 
up oxygen when dried in air. On heating the resin acids in a current of dry 
nitrogen, Herty and Dickson found that a small amount of carbon dioxide was 
given off at 65° -to 70°. 

It is evident from their work that the evolution of carbon dioxide at compara- 
tively low temperatures from turpentine and resin requires previous exposure to 
oxygen. Both of these substances are known to form organic peroxides when 
oxidized by the air. According to Engler and Weissberg,^^ pinene forms a peroxide 


of the formula CioHie<^ | , which on exposure to direct sunlight or warming to 


80° to 100° decomposes and gives up its "active** oxygen to "inner oxidation.** 
Thus turpentine yields, in addition to carbon dioxide, acetic acid, formic acid, 
a camphoric acid melting at 176°, and an acid isomeric with campholic acid.*"* 

Fahrion found that the peroxides formed by the autoxidation of colophony 
were quickly decomposed by heating to 100° and by long standing in strongly 
alkaline solutions. 

Dunlap and Shenk ^* and Genthe *° found that organic peroxides were formed 
during the autoxidation of linseed oil, and according to Mulder '" carbon dioxide, 
formic acid, and acetic acid are formed at the same time. 

The evolution of carbon dioxide from linseed oil, turpentine, colophony, 
the resin acids studied by Herty and Dickson, and Manila copal occurs 
simultaneously with the breaking down of the peroxides present. If 
oxygen is excluded and no peroxides are formed, no evolution of the 
gas occurs. The fonnation of carbon dioxide, under the conditions 
described above, can not be accounted for by the decomposition of a 
carboxyl-group in view of the nature of the substances from which it 
is evolved, since Herty and Dickson have shown that the resin acids 
in colophony are not decomposed at a temperature of 140°. Therefore, 
the conclusion seems warranted that the breaking down of the organic 

'^Journ. Ind. d Eng. Chem. (1909), 1, 68. 

"Bcr. d, deutachen chem. Qes. (1898), 31, 3046; (1900), 33, 1090. Compare 
Skraval, "Die induzierten Reaktionen,** 8amm, chem. u. chem.-tech. Vortrage. 
(1908), 13, 338. 

'^ Semmler, F. W., Die Atherischen Ole, Leipzig (1906), 2, 217. 

'*Joum. Am. Chem. 8oc. (1903), 25, 826. 

^"^Ztachr. f. ang. Chem. (1906), 25, 2087. 

^" Chemie der austrocknenden Oele, Berlin (1867), 109. 



222 BROOKS. 

peroxides in these substances causes the formation of carbon dioxide 
and probably formic, acetic/ and certain other acids also. According 
to Engler^s autoxidation theory the decomposition of an organic peroxide 
may cause the oxidation of another substance, the "acceptor/^ which is 
not autoxidizable. 

The air in several bottles containing powdered copal was washed through a 
solution of barium hydroxide and an abundant precipitation of barium carbonate 
was obtained. About 85 grams of powdered resin were placed in a tube and a 
stopcock sealed into each end. The air was washed out with pure oxj^gen, and 
on every day following about 200 cubic centimeters of moist oxygen were passed 
through the tube, dried by passing through calcium chloride, and the carbon 
dioxide absorbed in a potash bulb. The increase in weight averaged 0.017 gram 
per day. 

The amount of carbon dioxide given off from Mania copal at IOC* varies ac- 
cording to the previous treatment of the specimen. A sample weighing 5 grams 
was taken from the interior of a large lump, powdered, and heated to 100° in 
a "Current of hydrogen, which was then passed through a solution of barium 
hydroxide. Only a trace of carbon dioxide was detected after four hours heating. 
About 3.0 grams of a powdered sample which had stood for several weeks in a 
loosely stoppered bottle was treated in the same manner. The barium hydroxide 
solution became cloudy in a few minutes, and after four hours' heating 0.0162 
gram of barium carbonate was precipitated, equivalent to about 0.1 per cent of 
carbon dioxide. A portion of the powdered sample which gave no carbon dioxide 
was tested for peroxides with a solution of potassium iodide and starch with a 
negative result, while in the case of the specimen which yielded carbon dioxide 
a marked liberation of iodine occurred. These experiments were duplicated quali- 
tatively on several samples. 

The above phenomena could be caused by the terpenes in the copal, but 
terpene-free resin obtained by heating to 300° also showed the same 
behavior, which is consistent with the properties of colophony and 
abietic acid. Copal, free from terpenes, was also prepared by extracting 
an aqueous solution of the sodium salts of the resin acids four times 
with ether. The solution was then half evaporated on a steam bath to 
expel ether and the resin acids precipitated by dilute hydrochloric acid. 
The dry, powdered acids were exposed to air and sunlight in a covered 
crystallizing dish. All the changes noted as taking place in Manila 
copal occurred in this case, but much more slowly. Positive tests for 
formaldehyde and change of the photographic plate were easily obtained 
after four days exposure- to air and sunlight. 

' Schwalbe ^"^ suggested that the evolution of carbon dioxide from a 
' carboxyl-group during the process of saponification might affect the so- 
called "saponification^^ or Koettstorfer number. When phenolphthalein 
is used as an indicator in titrating the alkaline solution, the splitting 
off of carbon dioxide from a carboxylrgroup should have no effect on 
the Koettstorfer number obtained, since carbonic acid behaves as a 
monobasic acid to this indicator and strong alkalies. If the alkaline 

" Loc. cit. 



OXIDATION OF MANILA COPAL BY THE AIR. 



223 



solution resulting from a determination of the Koettstorfer number is 
acidified and the resin recovered, a second determination shows a decrease 
in the Koettstorfer number only when the sample of resin gives off 
carbon dioxide and shows a reaction for peroxides. 

Two series of Koettstorfer numbers were made on a portion of the sample which 

had been tested with negative results for the evolution of carbon dioxide and for 

peroxides. One series ot determinations was made in small Erlenmeyer ' flasks 

with open-air condensers. Another series was made in small Jena glass flasks 

of about 50 cubic centimeters capacity. A sample of resin Weighing 2.00 grams^ 

N 
10 cubic centinjeters of benzene, and 25 cubic centimeters of - alcoholic caustic 

potash were introduced into each of the Jena glass flasks, which were then sealed 
and heated in a bath of boiling alcohol in order that the temperature in the 
sealed flasks would not be higher than in the open ones^ The effect of prolonged 
digestion is also brought out by these series. In determining the Koettstorfer 
numbers of the recovered samples, the digestions lasted one-half hour each. 

Table I.^ — Koettstorfer numbers obtained under different conditions on a saynple 

of Manila copal. 



Time. 



24 hours at 30° 

One-half hour boiling . 

1 hour boiling 

2 hours boiling 

4 hours boiling 

6 hours boiling 

8 hours boiling 



Open'. 



144.6 
153.6 
162.9 
166.8 
170.6 
174.0 



Sealed. 



145.0 
145,0 
15H. 8 
160.0 
166.8 
170.4 
174.2 



Recovered 
resin. 



144.5 
154.6 
160.2 
166.8 



The recovered resins were dried in hydrogen to prevent oxidation. It 
will be noticed that their Koettstorfer numbers are the same as for the 
original samples. 

Powdered samples left exposed to the air for a few days show a 
peroxide reaction when tested with potassium iodide and starch. Such 
samples show a slight decrease in their Koettstorfer numbers after re- 
covering the resin from the first determination. This is to be expected, 
since part of the alkali is neutralized in the first determination by carbon 
dioxide and small amounts of volatile acids. The sample weighing 3 
grams, which yielded 0.1 per cent of carbon dioxide at 100°^ also gave 
up acids which were soluble in the water in contact with it, which 

N 
required 5.5 cubic centimeters of -^ caustic soda solution to make them 

neutral to phenolphthalein. The absence of these acids in the recovered 
resin would cause a decrease in the Koettstorfer number of about 10 units. 
A powdered sample which had stood in a loosely stoppered bottle for 
eight months showed a Koettstorfer number, after one-half hour's diges- 
tion, of 210. The recovered resin, dried in hydrogen, showed a Koett- 



224 BROOKS. 

storfer number of 187. This sample gave a positive test for peroxides 
and gave off carbon dioxide on heating to 100°. 

Henriques " made a series of similar experiments on colophony with the object 
of determining whether or not esters were present. The "acid numbers" 'of the 
recovered resins were in every case lower by 6 to 8 units than the values obtained 
for the original samples. 

The above table also brings out the well-known fact that prolonged 
digestion of resins in alcoholic potash yields high Koettstorfer numbers. 
According to Lewkowitsch ^^ abnormally high values may be obtained by 
prolonged digestion even in the cold. The series of determinations made 
in the small sealed flasks also show conclusively that the higher values 
are not caused by atmospheric oxidation during the course of the experi- 
ment, as was suggested by Worstall. Several determinations were made 
in an atmosphere of hydrogen with the same result. Two determinations 
were made on separate portions of the same sample, the only difference 
being that a current of hydrogen was passed through one flask and a 
stream of air, free from carbon dioxide, through the other. An increase 
of 15 units was observed in the Koettstorfer number of the latter and 
an increase of 13.5 units in the former case. 

Lewkowitsch says that the changes produced by prolonged digestion in alkalies 
point to "gradual degradation" of the resin. Fahrion -" and others have suggested 
that lactones may be present, which are formed by rearrangement of peroxides, 
similar to the reactions which Baeyer and Villiger ^^ found in the case of the 
peroxides of camphor, menthon, tetrahydrocarvon, and ketocyclohexane. Accord- 
ing to this theory the lactones present in the resins are slowly decomposed by 
alkalies and give rise to the gradually increasing Koettstorfer numbers. 

THE INCREASE IN THE KOETTSTOUFER NUMBER OF MANILA COPAL CAUSED 

BY AUTOXIDATION. 

A freshly powdered sample of Manila copal was found to have the 
Koettstorfer number 157. About 85 grams of this sample absorbed 135 
cubic centimeters of oxygen during the first fourteen hours after powder- 
ing. After standing exposed to the air for four months the Koettstorfer 
number, determined in the same manner, was 182. Another sample in- 
creased about 35 units during a period of eight months, having a final 
Koettstorfer number of 210. The differences observed in these cases are 
considerably greater than in the acid numbers recorded by Worstall. ^^ 
Although the so-called acid numbers and the Koettstorfer numbers are 
strictly comparable, it is probable that the generally higher temperature 

**Chem. CentraM. (1899), 2, 151; Chem. Rev. d. Fett. u. Harz-Ind. (1899), 
6, 106. 

^* Analyst (1901), 26, 37. 

^'ZUchr. f. ang, Chem. (1907), 20, 356. 

^Ber, d, deutschen chem. Ges. (1899), 32, 3625. 

* toe. eit 



OXIDATION OF MANILA COPAL BY THE AIR. 226 

prevailing here and the more intense sunlight of the Tropics would cause 
more rapid oxidation. 

The increase in the Koettstorfer number of a resin which has been exposed to 
the air also points to the rearrangement of peroxides to lactones. A similar 
phenomenon perhaps occurs in the case of linseed oil. Lewkowitsch ^* gives the 
saponification number of a sample of blown linseed oil as 287, or about 100 
units higher than for ordinary raw oil. 

Certain hydroxy-acids in the aromatic series give Koettstorfer numbers which 
are considerably higher than the values calculated for the carboxyl-group alone." 
This fact might be cited to show that the increase in the Koettstorfer numbers 
of samples of oxidized copal is due to the formation of hydroxy-acids, which 
might also form lactones. However, such an explanation would be difficult to 
apply to turpentine. 

Since the oxidation by the air of these widely different substances is 
evidently due to their unsaturated character, is, accompanied by the 
formation of oi'ganic peroxides^ and gives rise to reaction products the 
formation of which is readily explained by the breaking down of the 
peroxides, it is evident that the only explanation which will suit all of 
these cases is that the changes which take place during autoxidation 
depend on the formation of organic peroxides. 

OTHER PKODUCTS OF THE OXIDATION OF MANILA COPAL BY THE AIR. 

Formic acid has repeatedly been found as a product of the atmospheric 
oxidation of linseed oil and turpentine. It is also formed during the 
autoxidation of Manila copal. 

A freshly powdered sample of the resin possesses a pleasant aromatic smell, 
but on standing for a few days in a loosely stoppered bottle acquires a sharp 
irritating odor. A small evaporating dish containing a dilute solution of hydrox- 
ide was placed in a jar containing about 300 grams of powdered resin and 
allowed to stand for three weeks. The solution was then diluted, acidified with 
sulphuric acid, and distilled until one-half of it had passed over. This was 
neutralized by ammonia, concentrated to about 2 cubic centimeters, and the 
presence of formic acid shown by the ferric chloride test. Its presence was also 
shown by dissolving 200 grams of powdered resin in cold, diluted alkali, precip- 
itating the resin acids and filtering. The filtrate was then neutralized and 
concentrated, acidified with phosphoric acid, and distilled until about one-half had 
passed over. The distillate was treated as before and a positive test for forn\ic 
acid was obtained. 

Formaldehyde is also a product of the autoxidation of Manila copal. 

An open dish containing 10 cubic centimeters of water was placed in a jar 
containing 200 grams of powdered resin and allowed to stand for one -week. At 
the end of this time the water was strongly acid, and three drops were sufficient 
to give an unmistakable test for formaldehyde when tested by the well-known 
color reaction with milk and sulphuric acid containing a trace of ferric chloride 
and by the reaction between hexamethylene tetramine and bromine water. 

^^ Analyst (1902), 27, 140. 

^^Fahrion, Ztschr. f, ang. Chem. (1901), 14, 1226. 



226 BROOKS. 

Although formic acid has frequently been identified among the products 
of oxidation in the sunlight of many organic substances, formaldehyde has 
often been overlooked. As mentioned above, copal carefully freed from 
terpen es yields a small amount of formaldehyde on autoxidation. I 
have shown that formaldehyde is also formed during the autoxidation of 
pinene, limonehe, and linseed oil. In fact, the formic acid present may 
have been formed from formaldehyde, since, according to Kastle and 
Lovenhart,^*^ the oxidation of formaldehyde is accelerated by sunlight. 

A sample of pinene boiling between 155° and 157° was carefully purified and 
distilled over sodium. The oil w^as then exposed to moist air and sunlight for 
three days, after which time the oil was shaken out with an equal volume of 
water. Three drops of the aqueous solution were sufficient to obtain a distinct 
reaction for formaldehyde when tested by the milk and sulphuric acid reaction. 
A sample of limonene treated in the same manner yielded formaldehyde. 

Linseed oil, which had been exposed to air and sunlight for three days, was 
extracted with An equal volume of water and the extract tested for formaldehyde. 
Positive tests were obtained although the concentration of the aldehyde solution 
was much lo\Yer than in the previous cases. 

THE INFLUENCE OF SUNLIGHT ON THE ABSORPTION OF OXYGEN BY 

MANILA COPAL. 

Sunlight, so far as it has been tested in this respect, has a marked 
accelerating effect on autoxidation. 

Genthe^* showed that light, particularly of the shorter wave lengths, greatly 
accelerated the autoxidation of linseed oil. Light is also known to accelerate the 
autoxidation of turpentine. Klason and Kohler ^^ state, of the resin acids of the 
Norway spruce which they studied, that "The sensibility of the resin acids to 
oxidation is very different in winter and summer. It is almost impossible to work 
with sapin acid from May to August. It is evidently not the temperature alone 
which causes this. Either the light or a yet unknown cause is active here." 

Sunlight accelerates the oxidation of Manila copal by the air, as was 
shown by a simple experiment in which equal portions of a powdered 
sample were placed in two bottles of the same capacity and filled with 
oxygen. 

Manometers were attached to the bottles. One bottle was covered with four 
layers of tin foil and both Vere then exposed to direct sunlight. In five hours 
the diminished pressure in the unprotected bottle was equivalent to 30 centimeters 
of mercury while the manometer on the protected bottle showed a difference in 
level of only 14 centimeters. The difference in the temperature in the two bottles 
did not. exceed three degrees. After four days both manometers showed a dif- 
ference in level of about 43 centimeters. 

"^Journ. Am. Ghent. 8oc. (1899), 21, 268. 

^'^ Loc. cit. 

" Journ. f.prakt.Chem. {1906), 'N.F. 181,337. 



OXIDATION OF MANILA COPAL BY THE AIR. 227 

SUMMARY. 

1. Manila copal rapidly absorbs oxygen from the air. The oxidation 
is accompanied by the formation of organic peroxides, an increase in the 
Koettstorfer number, and evolution of small quantities of carbon dioxide, 
formaldehyde, formic acid, and hydrogen peroxide. 

2. The resin acids of Manila copal, when separated from the terpenes 
occurring in the natural resin, undergo oxidation by the air. 

3. The evolution of carbon dioxide from Manila copal and colophony 
is probably due to the formation of organic peroxides and their subsequent 
decomposition. 

4. The increase in the Koettstorfer number obtained by prolonged 
digestion with alcoholic potassium hydroxide is not due to oxidation 
during the course of the experiment, but is probably caused, at least in 
part, by the breaking down of lactones and organic peroxides. Samples 
which have been exposed to the air give up carbon dioxide and formic 
acid to the alkaline solution in the Koettstorfer determination and 
cause the recovered resin to show lower Koettstorfer numbers than the 
initial values. 

5. Formaldehyde has heretofore not been noted among the products 
of the oxidation of organic substances by the air. I have found it among 
the products of the atmospheric oxidation of Manila copal. 

6. The oxidation of Manila copal by the air is accelerated by sunlight. 



THE OLEORESIN OF PINUS INSULARIS ENDL. 



By Benjamin T. Brooks. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. /.) 



The preparation of turpentine and rosin, or colophony, from the resin 
of Pinus insularis Endl. is apparently a commercial possibility. It was 
therefore investigated in order to ascertain if the products compare favor- 
ably with the turpentine and naval stores manufactured in the southern 
United States. 

To judge from the work so far as it is finished, the oleoresin is identical 
with that collected in America. The volatile oil consists for the greater 
part of ordinary pinene and the resin of over 90 per cent abietic acid. 

The constants of the oil as determined by Mr. George F. Richmond of 

this laboratory are as follows:^ Specific gravity, ^^^=0.8593; N^= 

30 
1.4656; A^=-f ^6.0. Ninety-six per cent of. the oil distilled between 

lo4° and 165°.5. 

T have determined the optical rotations of Hwe specimens of the tur- 
pentine which were distilled with steam from five samples of the oleo- 
resin collected from as many different trees. Considerable variation in 
these values was found, but the differences were not as large as those 
noted in the case of American turpentine by Herty.^ The values I 
obtained at 30°, using a Schmidt and Haensch half-shadow polariscope, 
sodium flame, range from +13.15° to +27.48°. These numbers are 
sufficient to show that the optical rotation of the turpentine from this 
species, like that obtained from Pinus palustris Mill, and Pinus hetero- 
phylla (Ell.) Sudworth in America and the commercial turpentine of 
France and Russia, has no characteristic value. 

Although the constants identify the oil as consisting principally of 
ordinary pinene, several derivatives of the latter were prepared for 
further proof. 

From a fraction boiling from 155'' to 156*' pinene nitrosyl chloride was 
prepared; it was identified microscopically, by its melting point (115<») arid by 

^This Journal, Sec. A (1909), 4, 232. 
^Journ. Am. Chem,. Soc. (1908), 30, 863. 

229 



230 BROOKS. 

converting it into pinene nitrolbenzylamine. The fraction boiling from 159° to 
161°, 36 per c^nt of the oil, was tested for /S pinene by the method of Wallach* 
and a small quantity of nopinic acid, melting point 121°, was obtained. The yield 
of nopinic acid was less than 1 per cent. 

The colophony remaining after the distillation of the turpentine consisted 
almost entirely of abietic acid. No other acid has been isolated from it. A 
sample of the colophony was crystallized once from dilute alcohol and a yield of 
90 per cent of crude crystalline abietic acid was obtained. I have found that 
the yield of crystalline abietic acid which it is possible to obtain depends very 
much upon whether the turpentine has been removed by steam or dry distillation. 
A sample of dark brown American colophony when treated in the same manner 
separated as an oil, which crystallized only after several days, standing in an 
ice box. 

When the original samples were collected, some of the clear, honey-yellow 
exudate was poured into glass vials and sealed from the air. In two hours 
crystallization had set in, the mass changing to a white crystalline mixture. 
On returning to the laboratory the crystals were isolated by spreading 30 grams 
of the mixture on a porous plate which was then placed in a desiccator filled 
with hydrogen. After standing one week the crystalline residue remaining on 
the plate weighed 17 grams. The substance was recrystallized twice from dilute 
alcohol and it then melted at 128*' to 131°. By repeated crystallization from 
dilute alcohol alone, a melting point higher than 129° to 131° could not be 
obtained, but by crystallizing from alcohol diluted with strong aqueous hydro- 
chloric acid, crystals were deposited which melted at 154° to 156° without 
further recrystallization. The angles of the latter crystals were measured under 
the microscope and proved to be identical with those given by Mach * for abietic 
acid. Several concordant combustions agreed fairly well with the numbers for 
abietic acid. 0.1980 gram substance gave 0.5690 gram CO2 and 0.1760 gram HoO. 

Calculated for CigHggOo. Found 
Per cent. ' per cent. 

C 79.16 78.40 

H 9.72 9.87 

Klason and Kohler ^ suggested that abietic acid itself did not occur in turpen- 
tine oleoresins, but was probably formed from another acid which was very sen- 
sitive to heat. Wiesner * states that the substance crystallizing from ordinary 
turpentine oleoresin is abietic acid. 

In the present state of our knowledge it is impossible to state exactly what 
abietic acid is.*^ The nomenclature used by difl*erent authors is by no means 
uniform, yet there is no doubt but that the high melting acids, obtained by 
crystallization from solutions containing hydrochloric acid, are isomers of the 
original low melting acids.® In view of the fact that the properties of the acid 
obtained by me by crystallizing from dilute alcohol and hydrochloric acid agree 
well with those of the abietic acid described by Mach, this name is given the 
preference. 

^ Ann. d, Ohem, [LieUg) (1908), 356, 228. 
^Monatsh. f. Chem. (1894), 15, 629. 
^Joum. f. prakt. Chem. (1906), N. F. 181, 337. 
• «Die Rohstoffe des Pflanzenreiches. 2d ed. Leipzig, (1900), 1, 216. 
'Bucher, Journ. Am. Chem. Soc. (1910), 32, 382. 
"Leskiewicz, Journ. f. prakt. Chem. (1910), N. F. 81, 411. 



OLEORESIN OF PINUS INSULARIS ENDI.. 231 

No evidence of a crystalline substance other than abietic acid was found. 
Twelve grams of the crude substance were dissolved in 30 cubic centimeters of 
dry ether and dry ammonia passed in until the solution was saturated with the 
gas. Nothing separated from the solution. 

The mucilagenous material left in the mother liquors after crystallizing the 
crude abietic acid was not investigated. 

The Koettstorfer .number of a "water white" sample of the colophony was 
170.2. The Liebermann-Storch reaction gave a brilliant purple the same as in the 
case of American colophony. 

The above data are sufficient to show that the turpentine and colophony 
from Pinits insularis Endl. are practically identical with those produced in 
America 



Order No. 407. 

THE BATAN DIALECT AS A MEMBER OF THE PHILIPPINE GROUP 

OF LANGUAGES. 

By Otto Scheerer. 

TOGETHER WITH 

"F" AND "V" IN PHILIPPINE LANGUAGES. 

By Carlos Everett Conant. 

141 pages. 

Price $0.80 United States currency. 



Order No. 408. 

A VOCABULARY OF THE IGOROT LANGUAGE AS SPOKEN BY THE 

BONTOC IGOROTS. 

By the Eev. Walter Clayton Clapp. 

Igorot-English and English-Igorot. 

/ Price $0.75 United States currency. 



Order No. 102. 



A MANUAL OF PHILIPPINE BIRDS. 

By EiCHARD McGregor. 

Part I.-412 pages ) 

Part IL-324 pages [ In two volumes. 

Descriptions of all known species of Philippine birds^ together with 
keys to species and genera. 

Price (for the two volumes) $4 United States currency. 



THE PHILIPPINE 

Journal of Science 

A. Chemical and Geological Sciences 
AND THE Industries 



Vol. Y JULY, 1910 No, 4 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES: THEIR 
'SUITABILITY FOR PAPER MAKING.' 



By George F. Richmond. 
XFrom the Cheynical Laboratory, Bureau of Science, Manila, P, I.) 



TNTKODUCTION^. 



Since concluding our survey of Philippine fibers and fibrous substances 
from the standpoint of their suitability for the manufacture of pulp and 
paper, I have had occasion to investigate certain selected species of bamboo 
from this and other Malayan regions and to consider in some detail 
the various factors connected with the introduction into this part of 
the world o'f an industry for making paper pulp from bamboo. The 
prospective investor needs particularly to know^ where he can manufacture 
at a profit, the conditions under which he will have to work, the quantity 
of the raw material available, the native substances for the production of 
chemicals, and the costs of freights, power, fuel, and water, as well 
as the natural facilities and advantages of the site. Experimental results 
on the prevailing species of Philippine bamboos are therefore given in 
this paper, including a brief resume of the causes leading to the recent 
rather serious inquiry concerning this interesting material as a possible 
new source of paper stock. 

The last three or four years have witnessed a general extension of the 
search for new materials wherever paper is made or used. This is 
evidenced by the large number of new vegetable products which have been 
investigated and reported upon in relation to their suitability for the 
manufacture of paper. 

^ Continued from This Journal, Sec. A (1907), 2, li2. 
96918 233 



234 RICHMOND. 

A. D. Little, official chemist of the American J?aper and Pulp As- 
sociation, has discussed this suhject at full lengtli.- 

He has emphasized the "growing scarcity of pulp wood, the continually rising 
price and longer haul with which paper makers using this material are now 
contending, and the coming competition of new and better stocks which even 
now can be produced more cheaply than any bleached wood fiber. 

"Wood, as a raw material, has proved so available, convenient, compact, easily 
handled, and heretofore so cheap, that we have been led to overlook or ignore 
the immense sources of other and better paper stocks which easily lie within our 
reach. • 

"We are not dealing with the perennial suggestions of visionaries who see 
a paper stock in everything which has a fiber, but are, instead, concerned with 
the serious proposals of capable technologists whose conclusions are based on 
careful study." 

While opinions differ as to . the growing scarcity of suitable pulp 
woods, it is safe to assume that spruce wood is becoming exhausted. 
Spruce forms at least two- thirds of all the wood converted into pulp 
at the present time. The supply of this material was formerly considered 
to be practically inexhaustible. 

The importance of searching for a suitable substitute for wood is 
realized when it is considered that the world's annual production of 
paper has reached 8,000,000 tons, with an estimated increase of 25 per 
cent every tien years. Six and a half million tons of this output are 
made from wood. Therefore, wood is the controlling factor in the paper 
world to-day, and influences which tend to increase its cost are the 
direct cause of the recent systematic search for a cheaper substitute. 

BAMBOO FOR PAPER l^ULP. 

The trend of recent investigation in the direction of bamboo fiber 
is especially noteworthy, and the serious consideration of this material 
at present is largely the outcome of the spasmodic attention which it 
has received since 1875, the date of its first introduction. This revival 
of interest in bamboo as a source of paper stock is largely due to the 
investigations of E. W. Sindall and William Riatt, two English paper 
manufacturers and pulp experts, who, independently of each other, have 
studied the material in British India. 

Sindall, in a report ^ to the government of India, sums up his con- 
clusions on the suitability of bamboo fiber for paper making and the 
practicability of its commercial use as follows : 

In point of texture and strength the pulp obtainable is of excellent quality. 
With bamboo costing three dollars and twenty-four cents per ton, the value of 
the raw material required for a ton of unbleached pulp works out at a reason- 
able figure. As it would take about 2i tons of bamboo to make 1 ton of paper 

^Chem. Eng, (1909), 7, 106. 

'Report on the Manufacturing of Paper and Paper Pulp in Burma. (1906.) 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 235 

pulp, the cost of the raw material for 1 ton of finished pulp is seven dollars 
and thirty cents. 

At the prices quoted above the manufacture of paper pulp for export seems 
to be a practical commercial problem. A ton of unbleached bamboo pulp could 
be produced for about twenty-six dollars and seventy-six cents, including manu- 
facturing costs, interest, and sundry charges. This cost supplemented by the 
freight to England and sundry dues, would be increased to thirty-six dollars 
and fifty cents, as the price delivered in London or Liverpool. Having regard 
to the quality of the pulp, a higher price would be realized, since wood pulp is 
valued at eight pounds to nine pounds sterling per ton (thirty-eight dollars and 
ninety-three cents to forty-three dollars and eighty cents gold). 

Acting upon Sindall's recommendation that a practical trial with bamboo on 
a large scale", would go far toward ultimately bringing his preliminary inquiry 
to a successful issue, several tons of bamboo were shipped to the English paper 
mill of Thomas & Green, at Woburn, near London, for a practical demonstration 
of its paper-making qualities. 

This firm reported that the material worked exceedingly well on the machine 
and produced a good sheet of strong paper suitable for special kinds of printing, 
particularly for lithographic work. A portion of the bleached pulp was sent 
to the North Ireland Paper Mill Company, whose manager reported as follows: 
"We found no difficulty whatever in working the stuff. It was put through our 
mill just the same as if we had been treating wood pulp, without any alterations 
to the machinery being necessary. The paper thus made was submitted to 
lithographic printers who reported that it took both letter-press and lithographic 
work admirably." 

The bamboo was cooked with caustic soda of 15° Twaddell for seven hours 
at 60 pounds pressure and bleached to a good white with 8 to 10 per cent of . 
bleaching powder. The complete results of this practical test were published in 
a pamphlet on paper made from the bamboo. 

A. D. Little's opinion * of Sindall's booklet is of value in this connection. He 
writes: "There appears to be no room for doubt that bamboo is one of the most 
promising, if not in fact actually the most attractive of the new sources of 
paper stock available at the present time." 

It would seem to me that the practical test quoted above is final and 
conclusive for the species of bamboo selected and the process of manu- 
facture employed, and that the last doubt of the most skeptical regarding 
the intrinsic value and applicability of bamboo fiber for certain grades 
of paper is entirely removed. 

Mr. Riatt's investigations on bamboo have extended over a period of ten 
years since 1897, during which time many fundamental questions regarding the 
commercial use of the material have been thoroughly considered. He writes:'^ 
"Among the fibrous products of our tropical and subtropical forests, none is more 
likely to take a more leading place as a paper-making material than bamboo. 
Its accessibility^, being generally found within reach of waterways down which 
it may be rafted; the size of the stems, giving a larger return per head per day 
for the cutting and collecting force employed than in case of the smaller annual 
grasses; the ease with which it yields to the same methods of treatment which 
have been so successful with wood, together with its greater abundance, mark it 

*The Paper Mill and Wood Pulp News (1910), 35, 37. 
""Trop. Agr. (1909), 6, 32. 



236 RICHMOND. 

out as the fittest and most ' natural successor to the position occupied by spruce 
and pine trees during the last thiry years; while its power of self-reproduction 
makes it impossible that the process of exhaustion of supply, which has taken 
place in the case of these timbers, can ever happen with bamboo.' 

The particular value of Eiatt's results lies in his extensive observations 
on the manner and habits of growth and reproduction of selected species 
of Indian bamboos and the applicability of the information obtained to 
the bamboos of other tropical or subtropical regions. This practical phase 
in the commercial use of bamboo will be taken up at another place, when 
the availability of the Philippine bamboo forests for commercial use is 
discussed. 

An extended laboratory study of a new or hitherto commercially 
untried raw material for a specific purpose should be of a comparative 
nature. In judging of the pulp and paper making qualities of bamboo, 
it should be directly compared with the material with which it will 
necessarily compete on the world's markets, namely, pulp wood. The 
comparison should consider, first, the applicability of present methods of 
manufacture, and, second, the suitability of the manufactured product 
to present uses. While I fully realize that a laboratory examination of 
a given raw material for a specific commercial use is not as conclusive 
to the average prospective investor as a practical test under actual man- 
ufacturing conditions, in the absence of an opportunity for such a practical 
trial I have endeavored to approximate actual factory procedure in all of 
my experimental work, and, personally, I am satisfied that the results 
obtained will not prove misleading. 

At the present time wood pulp is manufactured by two distinct proc- 
esses: (1) the so-called mechanical process, which consists in grinding 
the wood into pulp by means of rapidly revolving stones; and* (2) the 
chemical process, in which the wood is reduced to pulp by the direct 
action of alkalies or acids. 

The mechanical process is not applicable because of the physical 
structure of bamboo, hence the commercial utilization of this material 
for paper stock is limited to chemical means of pulp production. 

Until quite recently, it was considered that the acid or sulphite proc- 
ess was better adapted to certain species of woods than that of soda or 
alkalines and vice versa. In 1900 poplar was the only wood made into 
soda pulp, while in 1907 spruce, balsam, and some hard woods were 
prepared in this way; and, although statistics for 1907 show that over 95 
per cent of all the wood treated by the sulphite process was spruce and 
hemlock, this does not necessarily mean that it was employed because of 
the nature of the raw material, but rather because the governing condi- 
tions were such that it was cheaper of application. It remains to be 
seen whether the character of hard woods, none of which to my knowledge 
as yet have been subjected to the sulphite process, limits this method of 
manufacture. * 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 237 

All other raw or waste materials used in the industry, such as straw, 
grasses, rags, hemp, and jute, are almost invariably subjected to some 
alkaline method of treatment. Practical evidence that these materials 
are not amenable to the sulphite treatment is lg,cking. Theoretically, 
both processes are applicable and should produce their respective classes 
of fiber from the same fibrous vegetable material. 

It is well known that alkalies are more vigorous in their action upon 
vegetable structures than acids, and, therefore, are probably better adapted 
for those plants which are disintegrated with difficulty. It has always 
been believed that highly mineralized vegetable substances, such as cereal 
straws, would not lend themselves readily to the sulphite process. 

Bamboo is highly mineralized, and, like straw, possesses a very hard, 
impervious siliceous coating. To my mind, an experimental study of the 
degree of applicability of the present chemical processes of pulp man- 
ufacture to a new raw material should consider the physical structure of 
the plant in question as well as its chemical composition. 

One of the most common species of Philippine bamboos is a thin-walled 
variety known locally as cana ho jo, a small bamboo {Schizostachyum 
mncronatum Hack.). It grows in rather ill-defined clumps of 20 to 50 
stems each. The mature stems average at least 12 meters in height and 
6 centimeters in diameter at the base. The stem walls vary from 4 to 6 
millimeters in thickness throughout the entire length, and the distance 
between the joints or nodes averages about 80 centimeters, although this 
is much shorter near the base of the stem. The nodes must be removed 
for the sulphite process of treatment; for the soda process, I am inclined 
to think that the entire stem may be crushed and chipped, subsequent 
screening removing any uncooked portions. 

The sulphite process presents almost insurmountable difficulties in an 
experimental way, although it is well under control in its practical ap- 
plication. After -numerous interruptions due to leaky digesters and a 
water supply rendered unsuitable by heavy rains, I was able to produce 
variable but controlled conditions of treatment, thus obtaining sufficient 
data for the general conclusions which follow. These experiments were 
conducted approximately under factory conditions of treatment and gave 
data upon the following points: 

I. Method of preliminary preparation of the bamboo for the diges- 
tion. 

II. The conditions of strength of liquor, duration and temperature 
of treatment. 

III. Yield, bleach-consuming power, and cost of production. 

NATURE AND PRELIMINARY PREPARATION OF THE MATERIAL. 

The bamboo should be of uniform age and growth, that is, shoots of 
one, two, three, and four years^ growth should not be mixed indiscrimi- 
nately. Bamboo of different ages can equally well be treated separately 



238 RICHMOND. 

by varying the conditions of the digestion; the older growths requiring 
conditions of strength of liquor, temperature and time, which would 
result in considerable loss and the weakening of the fibers of younger 
ones. The best results were obtained with what I consider to be two- 
year-old bamboo, namely, shoots which had passed through two full 
seasons of growth. At this stage in the life of the plant the sheaths 
which formerly surrounded each node have fallen off and a crown of 
leaves has appeared at the top. Such a bamboo loses approximately 35 
per cent of its weight on being dried in the air and contains about 9 
per cent of nodes. It is sufficiently mature to give good yields of fiber, 
and it has not become too much lignified to be difficult to treat ; further- 
more, its removal at this stage will not cause any injury to the parent 
rootstock. I do not anticipate any trouble in securing bamboo of uniform 
age if one or more of the following precautions are observed : 

1. Young shoots of less than one year's gi'owth are very soft and 
tender. They have a troublesome sheath surrounding each node, but 
possess no leaves or branches. 

2. Bamboos two years old have lost their sheaths and a crown of 
leaves has appeared. 

3. Stems of three years' growth or more have a well-developed crown 
and branches extending halfway down the stem. 

All bamboos should be as completely air dry as possible in order to 
require the least time in cooking.^ If this precaution is observed they 
will crush and chip into a much finer state of division than they other- 
wise would. An air-dry cane passed between a pair of crushing rolls 
cracks or splits longitudinally into strips varying from 1.270 to 2.540 
centimeters in width, and when these strips are fed into an ordinary 
cutting box, such as is employed for cutting corn stalks or other fodder, 
the chips separate, more or less completely, still further into pieces from 
1.270 to 0.318 centimeters in width. Bamboo forms much denser indi- 
vidual pieces than w^ood and a subsequent crushing in a special chip 
crusher would prove very advantageous. No such apparatus was at 
hand in the laboratory and therefore all preliminary preparation was 
made as described above. Of course, in practice, the length of the chips 
is regulated more or less at will. 

I would recomend as machinery to prepare bamboo, sets of crushing 
rolls of the necessary capacity and speed, and especially constructed 
chippers of heavier construction than the ordinary rope or rag cutters, 
but not nearly as strong and massive as pulp wood chippers. 

" It will probably be necessary to construct cheap drying sheds in the mill 
yard for the storage Of the material during the rainy months of the year. 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 



239 



EXPERIMENTAL. 
DIGESTION OF THE PREPARED BAMBOO CHIPS BY THE SULPHITE PROCESS. 

Bamboo chips prepared as described above, in lengths varying from 
1.270 to 2.540 centirneters, but uncrushed, were well screened from dust 
and dirt and packed into a stationary, upright, lead-lined digester and 
heated with direct fire in the presence of sulphite liquors of different 
concentration and under varying conditions of temperature and time. 
Thirty-seven separate digestions were made, but in no instance was I 
able to produce from bamboo a pulp easy to bleach with bleaching powder, 
the universal bleaching agent employed in the industry at the present 
time. 

The process yields fully 50 per cent of unbleached pulp and with a 
much lower sulphur consumption than is required in commercial prac- 
tice for wood. Well prepared, but uncrushed chips pulped readily with 
liquors of ordinary* strength in six to eight hours, but the unbleached 
fiber was not as light in color as sulphite spruce and gould only be used 
in the unbleached condition for wrappings, tags, etc., where strength, 
rather than color, is the important consideration. It is needless to say 
that I varied all the conditions of the treatments in every conceivable 
manner with the main point in view of producing a pulp which would 
bleach readily, and with a reasonable consumption of bleaching agent, 
but without success. If bamboo pulp were most suited for use i-n an 
unbleached state, then the sulphite process should be adopted by all 
means, but the material is not sufficiently light in color to be mixed 
with mechanical wood pulp in preparing news print paper, besides it is 
too good a fiber for the latter or for wrappings, for which it is entirely 
suited so far as color is concerned. In my opinion, bamboo fiber is 
eminently fitted for paper for books and for certain grades of writing 
and lithographic papers, either alone or when blended with rag or sul- 
phite wood pulp. 

A few data selected from the more successful sulphite experiments 
are given: 

Table I. 



Experi- 
ment No. 


Composition of the liquor. 


Conditions of time 
and temperature. 




a 
o 
w 

3 


1 . 

a pk 

it 

o o 

02 


Color. 


0) 

3 


'6 

i 


73 

o 


'6 

'I 


III 


6 

a 

1 


Is 
wag. 


1 


Per ct. 
1.17 
1.28 
1.09 
1.12 


Perct. 
3.87 
4.53 
3.39 
3.71 


Perct. 
1,34 
1.47 
1.25 
1.28 


Perct. 
2.52 
3.04 
2.14 
2.43 


Hrs. 
H 
3 
4 
3 


Hrs. 

7 
7 
9 
8 


146 
150 
145 
145 


Perct. 
52.4 
50.2 
50 
61 


Perct. 

80 

27.6 

24.12 

28.5 


Metric. 
149.6 
240.0 
240.0 
223.5 


Poor white. 
Do. 
Do. 


2 


3 


4 





240 . RICHMOND. 

It will be noted that from 24 to 30 per cent of bleaching powdef was required 
to produce at best a poor white, that is, 12 to 15 kilos of bleaching powder of 
the standard strength (35 per cent available chlorine) are required for 50 kibs 
of unbleached pulp. This consumption is excessive, and it forms the greatest 
objection to placing a bamboo pulp prepared in this way on the market. Esparto 
requires 5 to 7.6 kilos and wood 6 to 12.5 kilos. The consumption of sulphur 
is low and no account was taken of a certain percentage of recovery which is 
always possible in practice. In sulphite wood pulp manufacture, 150 to 175 
kilos of sulphur are usually required for the production of a ton of pulp, although 
recent attempts at more efficient gas recovery have brought the sulphur con- 
sumption per ton of pulp down to 115 kilos in some cases. 

The time required to produce an apparently well-cooked pulp was from six to 
eight hours. This is in keeping with modern tendencies, and a rather weak 
liquor (No. 3) produced as good a yield and a rather better bleaching pulp than 
did a strong liquor (No. 2). 

Aside from the poor bleaching properties of bamboo siilphite fiber 
prepared under the above conditions, there are other factors, both local 
and general, which tend to preclude the use of the sulphite process of 
treating bamboo at the present time. 

1. Bamboo fiber appears better suited for book printing and litho- 
graphic papers than for wrapping or news printing paper. This being 
the case, bulk, softness, and opacity, which are the chief features of soda 
fiber, are what is desired. 

While sulphite fiber also enters into the composition of some" grades of book 
paper; this is probably due more to a question of the cost of production, or the 
applicability of the raw material to the process, than to any inherent qualities 
of the fiber itself for the particular purposes. Sulphite fiber produces the 
stronger, harder, and more transparent papers demanded for tags, labels, wrap- 
ping and news printing, and the cost of soda fiber is such that it is particularly 
suited for blending with sulphite wood pulp, or with rags to produce the better 
class of book printing and even certain commercial grades of writing papers. 

2. It is undoubtedly true that the sulphite process costs less than the 
other for chemicals. Sulphur, at present quotations, can be convoi'ted 
into sulphite liquor and thrown away after use at less expense than the 
cost of soda actually consumed plus the cost of its recovery. However, 
the local supplies of limestone are better adapted for making soda than 
sulphite liquor. 

While a straight limestone can be, and is, used to some extent for 
making sulphite liquor, a dolomite, that is, a stone carrying a high 
percentage of magnesia, is much preferred. On the other hand, a good 
lime for causticizing soda liquors should be as low in'magnesia as pos- 
sible. Thus far no dolomites have been found in the Philippines. 

The other cliemicals, sulphur on the one hand and soda on the other, 
used in the two processes, are not produced locally, hence would hate to 
be imported from the most favorable foreign source. 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 241 

3. According to Little/ the modern tendency in operating the sulpliite 
process is more toward the use of stronger sulphite liquors than formerly, 
particularly in Sweden, where liquors carrying 5 per cent or more of 
total sulphurous acid are being employed to advantage, the quantity of 
pulp being improved and a better recovery of gas obtained. There is 
also a tendency toward a shorter time of cooking and there is no longer 
any doubt but that pulp of first quality 'can be produced in seven hours, 
or even less. Of course^ this meahs increased output from the same plant. 
It will prove to be difficult to make the stronger liquors and to maintain 
their strength under tropical conditions of temperature, and this fact 
will place any attempt to manufacture sulphite pulp here at a disadvantage 
in competition with pulp manufacture in temperate regions. Even in 
the United States of America and in Canada, considerable trouble in this 
respect is encountered during the summer months. Only one material 
advantage would be possessed by pulp manufacturers under the prevailing 
local conditions of temperature ? and this could be a smaller loss of heat 
by radiation^ but this gain of course applies to any chemical process of 
cooking fibers with steam. During the winter months of the year this loss 
is a very material item of fuel expense in temperate climates. High 
temperatures have no deleterious effect on the liquors used in the soda 
process. 

4. The initial expense of installation ol the two processes, sulphite 
and soda, is approximately the same if the most modem equipment is 
purchased. If there is any difference it is in favor of the soda process. 

5. The cost of maintenance in other places is considered to be about 
equal between the two, but in the Philippines I am inclined to think that 
it would be less for the soda process : First,- because this method does 
not demand quite as many skilled operators, and, second, the sulphite 
digester linings are a constant source of trouble, whereas the improved 
form of welded soda digesters leave nothing to be desired in this respect. 

THE SODA PROCESS ON BAMBOO. 

It was found as the result of repeated trials with caustic soda liquors 
under varying conditions of strength, pressure and duration of cooking, 
that bamboo chips prepared as outlined above invariably yielded 43 to 
45 per cent of air-dr}' , unbleached fiber under the following conditions : 

(a) Upright cylindrical stationary digesters. 
(h) Direct live-steam heat. 

(c) Fifteen to 20 per cent of 76 per cent ^caustic soda calculated on the 
air-dry weight of the raw material. 

(d) A duration of cooking of four to six hours. 

(e) A maximum temperature 160° C. (320° F.) corresponding to a steam 
pressure of 45 kilos (90 pounds). 

'Paper Trade Journal (1908), 46, 101. 



242 RICHMOND. 

Fiber thus prepared bleached to a splendid white with 12 to 15 per 
cent of bleaching powder. The fiber was strong, of good felting capacity, 
and it made a more bulky sheet than wood pulp. Bamboo fibers average 
2.5 to 3 millimeters in length, so that they are somewhat longer and 
materially narrower than spruce fibers. 

That bamboo is readily resolved by the soda process of treatment to 
a fiber which is easily blended has been proved beyond doubt, and further 
experiment in this direction is scarcelj^ necessary. The fiber possesses 
the requisite lengthy strength, and felting capacity to meet the paper 
maker's demands, and the quantity of resistant cellulose per unit weight 
of the raw material is sufficient to warrant its extraction. Therefore, 
those factors other than raw material which control the economic use of 
bamboo for the purpose in question alone remain for discussion. The 
whole problem resolves itself into determining the cost of placing a ton 
of the finished product upon the local or foreign market. 

COST OF THE RAW MATERIAL. 

While a considerable amount of batnboo is annually removed from 
the public forest for various purposes, it is not taxed at the present 
time, hence no statistics of the quantity thus consumed are at hand. 
Furthermore, the unsystematic state of the present industry* affords 
no reliable basis for estimating the cost of collecting large quantities 
of the material under well-organized conditions. The price at which 
bamboo can be obtained in the open market and the items of cost which 
go to make the present market rate are for the moment the only reliable 
source of information. Bamboo can be obtained in Orani, Bataan Prov- 
ince, at from 8 to 12 pesos per 1,000 stems, averaging 7.5 meters in length, 
according to the season. 

The present method of cutting and transporting the bamboo from the public 
forests to the Orani market may be taken as in use elsewhere throughout the 
Philippines. The bamboo stems are chopped off with a heavy knife about 1 
meter from the ground, trimmed into lengths of 7.5 meters, and bound into 
compact bundles of 50 stems each for convenience in hauling. 

The bamboos are cut, trimmed and bundled at the rate of 30 to 40 centavos 
per 100 stems. The bundles are hauled through the forest for a distance of 1 to 
2 kilometers to the Colo River and then rafted down the stream's to 5 kilometers 
to Orani. 

An average day's work for a man and carabao consists in hauling 6 bundles 
or 300 stems from the forest to the river at a cost of 1 peso to 1 peso and 25 
centavos.* The labor cost, then, of cutting and transporting 1,000 stems to the 
river varies from 6 pesos and' 33 centavos to 8 pesos and 16 centavos. The 
difference between these figures and the price per 1,000 stems on the Orani 
market represents the labor cost of rafting and the dealer's profit. 

®One peso is equal to fifty cents United States currency. 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 243 

It is to be noted that the above figures refer to the limited and un- 
organized industry of supplying the seacoast towns with selected bamboo 
for building purposes. It will readily be seen that present costs are 
excessive and of no value except for illustration and comparison and 
they may be reduced by improved practices. At the same labor cost of 
cutting the yield per unit area may be considerably increased by em- 
ploying stems of greater length. Over e50 per cent of the present cost 
is consumed in land transportation by an obsolete method which would 
largely be eliminated by the use of modern methods of conveying the 
bamboo to a central factory situated near the place where it grows where 
good, clear water for manufacturing purposes is available. 

QUANTITY OF BAMBOO AVAILABLE. 

An experimental cutting for the purpose of determining the yield of 
commercial bamboo per unit area was made under the supervision of the 
Philippine Bureau of Forestry, as follows : 

, Location 5 kilometers (3.1 miles) southwest of Lamay, Bataan, 3.75 kilometers 
(2 miles) in a direct line from the seacoast. 

Date cut, September 17, 1909. 

Elevation, 80 meters (250 feet). » 

Area of plot, 30 by 33.3 meters (0.1 hectare). 

Number of stems cut from the plot, 1,075. 

Length of the weighed stems, 7,5 meters (25 feet). 

Green weight of l,(h5 stems, 6.181 metric tons. 

Air-dry weight of 1,015 stems,, 3.524 metric tons. 

The plot selected was fairly level, which facilitated the cutting; in other 
respects it was considered to be representative of the average Stand of commercial 
bamboo over an indefinite area in the immediate vicinity. The plot was cut clean 
of all standing bamboo, each stem was severed about 0.5 meter from the ground, 
trimmed into lengths of 7.5 meters and then cut into shorter sections for con- 
venience in weighing. 

Later, on October 25, the plot was again visited and 12 bamboos over 9 meters 
(30 feet) in length were cut, sawed into short sections, and weighed. They were 
then crushed, baled, and brought to the laboratory for a determination of the 
air-dry weight and the percentage of nodes. The green weight of 12 stems was 
86.36 kilos, or 7.2 kilos per stem; the air-dry weight was 49 kilos, or 4 kilos 
per stem ; the air-dry weight, minus the nodes, was 45 kilos, or 3.75 kilos per stem. 

Applying the data thus obtained to the figures of the yield on the 
experimental plot, it is seen that 0.1 \iectare would yield fully 4 metric 
tons of air-dry material free from nodes, and that this weight is rep- 
resented by 1,075 stems, or the approximate unit of quantity upon which 
the present labor costs of cutting and handling are based. Four metric 
tons of bamboo will produce approximately 2 short tons of pulp. Even 
at 10 pesos per thousand stems, which is the average market price in 
Orani, the cost of sufficient bamboo to make a ton of pulp is at the very 



244 RICHMOND. 

attractive figure of 2.50 pesos, as compared with about 30 pesos, as the 
present cost of 2 cords of pulp wood, the amount required to make a ton 
of wood pulp. 

In order to determine if there is a sufficient area of bamboo in the 
Philippines favorably located with respect to fuel and other manufactur- 
ing facilities to meet the demands of a pulp mill of moderate capacity, I 
traversed nearly 100 kilometers of the forest patrol trails of Bataan 
Province in October, 1909, and I can state positively that there are a 
number of areas of bamboo of uniform stand within 5 kilometers of 
the seacoast, any one of which would supply a pulp mill of 20 tons 
daily capacity indefinitely. In every case a stream of good, clear water 
is close at hand and wood for fuel is both cheap and abundant. 

An area of 1,000 hectares of bamboo would supply a pulp mill of 20 
tons daily capacity for three years of 300 working days each, not taking 
into account the growth of the bamboo during this period ; or, working on 
a three-year rotation, which is entirely feasible, such an area would supply 
a mill of this capacity indefinitely. Since an area of 2 kilometers radius 
contains over 1,200 hectares, it is seen that the question, of transportation 
of the material to a given point, so far as the distance of the hauling is 
concerned, is n©t a serious one. 

COST OF MANUFACTURE. 

The cost of pulp production, exclusive of that of the raw material, 
may conveniently be divided into the following items: 

(1) labor; (2) fuel, power, etc.; (3) chemicals and supplies; (4) repairs, 
renewals, depreciation; (5) taxes, insurance, interest. 

These items of cost will be discussed in some detail in order to show 
the effect of existing local conditions and to indicate wherein, if any, 
advantages exist in favor of pulp manufacture in the Philippines. 



I believe a distinct advantage will eventually result in this item in 
favor of the Philippines, although this will be slight in the initial stages 
of the industry. 

Pulp manufacture requires skilled labor in its numerous departments, 
for the quality of the finished product calls for close attention to details 
from the time the raw material is landed in the mill yard until the 
finished product is packed and baled for shipment. Certain phases of 
the work, such as cooking the material and attending to the machinery, 
will always demand the services of operatives brought up in the industry. 
The preparation of the raw material, the washing and screening of the 
stock and the making of the liqudr can no doubt eventually be conducted 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 245 

by local labor; and the power and repair plants can be operated by 
trained Orientals from the start. 

Pulp-mill workers may be obtained from a number of sources. Grenerally speak- 
ing, Americans and Canadians receive double the wages of individuals of the 
same class in Europe. Such being the case, it would be advisable to depend upon 
skilled labor from Europe, and it is safe to assume that practically double the 
wage scale obtaining in the home land would be demanded. This would place the 
labor cost of pulp manufacture in the Philippines on a " par with that in the 
United States and Canada, but at a disadvantage with respect to Europe. This 
would distinctly be noticeable when it came to invading the European market with 
the manufactured product. 

However, there are numerous occupations connected with pulp manufacture, 
such as plumbing, carpentry, blacksmithing, saw filing, knife grinding, masonry, 
firing, etc., for which satisfactory Oriental labor is available; furthermore, the 
demand for semi- or unskilled labor in a pulp mill is great arid herein lies what- 
ever advantage the manufacture of pulp in the Philippines would possess over 
Occidental localities. 

The following estimate refers to the industry from the time the bamboo is 
delivered at factory site until the air-dry unbleached pulp is baled for shipment; 
it is therefore exclusive of all costs of collecting and transporting the bamboo from 
the forest to the mill, which latter is included in the expense of the raw material. 

I. In the yard. — Including all handling, sorting, piling and conveying the bam- 
boo from the yard to the mill; sixteen Filipino laborers. 

II. In the preparing room. — One head European preparer in charge of sawing, 
chipping, crushing and screening the material; six Filipino artisans. 

III. In the liquor-making plant. — Two European alkali makers; two European 
evaporator men; six Filipino helpers. 

IV. In the digester room. — Two European head and two second cooks; six Fili- 
pino helpers. 

V. In the pulp-machine room. — Two European machine tenders; four helpers of 
the best grade of local labor; four laborers. 

' VI. In the power plant. — One head and one second engineer; two head firemen; 
six laborers, all Filipinos. 

VII. In the repair plant. — One machinist and electrician; six artisans, all 
Filipinos. 

Summary. 

Pesos. 
10 employees, exclusive of superintendence and mill foreman, 

at 6 pesos per day 60.00 

4 employees (best grade of Filipino mechanics), at 4 pesos 

per day 16.00 

18 employees (Filipino artisans), at 2 pesos per day 36.00 

38 employees (Filipino common laborers), at 1 peso per day.. 38.00 
Superintendence, stenographer, paying and shipping clerks.... 30.00 

Total -... 180.00 

At least 10 skilled pulp-mill workers, that is, alkali makers, recovery plant men, 
cooks, and pulp-machine tenders will need to be obtained from abroad. This 
number is necessary because their respective departments operate continuously in 
two shifts of eleven and thirteen hours, respectively. At least 18 local artisans 



246 RICHMOND. 

will be required to supply the demands of the power and repair plants and 38 
unskilled laborers will be necessary for handling materials and supplies, washing 
and screening stock and as helpers in the general routine of a pulp mill. 

The estimated cost of the different grades of labor required is considered by 
me to be very conservative. The amount given is considerably more than double 
the average wage of European pulp workers, and 2 pesos and 1 peso for skilled and 
unskilled labor is in excess of the prices of such labor in other local industries at 
the present time. Based on a daily production of 20 tons, which is approximately 
the minimum of economic production in the industry, the estimated daily cost of 
labor is only 9 pesos per ton of output, which is a very favorable expenditure as 
compared with that in the United States and Canada, where there are larger units 
for pulp production operated in conjunction with paper manufacture, but with a 
labor cost equal to the above in many of the departments. 



The cost of power, next to the cost of labor, probably enters into the 
total expense of manufacture to a greater extent than any other element. 

The making of chemical pulp requires steam for heating the digesters 
and for drying cylinders of the pulp machine, so that only under very 
favorable circumstances is the initial cost of a water-power installation 
warranted. 

The heavy power consumption in hauling, barking, sawing and chipping wood 
for chemical pulp does not enter into the estimates of the requirements of a 
plant utilizing bamboo. The material is light and is easily handled, there is 
no bark to be removed, and comparatively light saws, crushing rolls and chippers 
are all that are required to place this raw material in a condition suitable 
for the digesters. I estimate that 25 horsepower will perform all the work of 
the bamboo-preparing plant as compared with 100 or more horsepower in the 
wood room of a chemical wood-pulp mill of the same capacity. 

There is much to be said in favor of water power in respect to the low cost 
of maintenance and the small operating expenses, but in this country the incon- 
venience of locating the plant with due regard for the necessary capacity for 
water storage and the evident variations in water level because of climatic 
changes, argue against its economic value. 



Imported coals, either Australian or Japanese, form the bulk of the 
fuel employed by the local industries at the present time. 

The present quotations on these coals are 9 pesos and 50 centavos to 10 pesos 
and 50 centavos, ex-ship Manila, with an import duty of 50 centavos per ton. At 
least one local source for coal may be considered to be fairly well developed. The 
Bataan Coal Company's mines, situated about sixty hours' steaming distance from 
■ Manila, are at present turning out 300 tons of coal each day, and loading it on 
board ship at the rate of 6 pesos and 50 centavog per ton. A lower price than 
this might be expected by the steady consumer. The capacity of this mine is now 
being increased to 1,000 tons per day. 

The comparative steam-making value of Philippine and foreign ooals offered 
on this market, together with recommendations on the type of fire-box grates 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 247 

and method of firing the local coal, have already been published by A. J. Cox.'* 
of the Bureau of Science. 

In Bataan Province, wood, as a fuel, appears to be both cheap and abundant, 
if it is considered in connection with the timber industry, which is already well 
developed. Waste materials from the timbering operations, such as, slabs and 
the limbs of hard resinous woods, which now for the greater part are allowed 
to rot, could be used, and, in addition, the standing timber of the bamboo areaa 
and the mangrove along the seacoast are available for fuel. 

CHEMICALS. 

The chief argument against the introduction of a paper-pulp industry 
in the Orient has been the cost of imported chemicals. It is true that 
in this respect a local industry would be at a disadvantage in comparison 
with the leading pulp-producing countries. The United States, Canada, 
and Europe possess well-developed industries for quarrying and burning 
limestone, and it is safe to say that the production of lime, the material 
so universally employed in the chemical pulp industry, is economically 
developed in those countries to a maximum degree. Limestone or marine- 
shell deposits are quarried and burned in the Philippines in sufficient 
amount to supply the present demands, but the industry is practically 
undeveloped. 

A 20-ton soda pulp mill will annually consume approximately 2,000 tons of 
well-burned lime. To produce only this limited quantity would require a plaftt 
of comparatively small output, which would be run spasmodically, and although 
the advantages of cheap fuel and labor would probably partially offset these 
defects, it can not be expected that lime could be produced here for less than 
the maximum cost of manufacture in other countries. 

The cost of lime manufacture^^ in the United States varies from 
one dollar and twenty cents to two dollars and ninety cents per short ton. 
This corresponds to from 4.2 to 10.15 cents for 35 kilos (per bushel 
of 70 pounds). 

The cost per ton is divided as follows: 

Cents, U. S. 
currency. 

Interest on the cost of plant and quarry 5-20 

Taxes and minor supplies A 10-15 

Cost of quarrying 2 tons of stone 50-90 

Cost of fuel for burning 2 tons of stone 30-75 

Cost of labor exclusive of quarrymen 20-30 

The minimum estimate of $1.20 United States currency might be 
attained by a good modern plant running steadily under exceptionally 
favorable conditions as regards quarrying, fuel, and labor. The maximum 

""This Journal, Sec. A (1908), 3, 301. 
^"The Mineral Ind. (1906), 15, 552. 



248 RICHMOND. 

of $2.90 could easily be exceeded in a small plant or one operated 
unsteadily.^^ 

Caustic soda, soda ash or sodium sulphite, according to the modifica- 
tion of the process which was to be adopted, would need to be imported 
from Europe or America. Based upon the present quotations and al- 
lowing the very conservative estimate of ten 'dollars per ton for freight 
from New York via Suez, caustic, soda, c. i. f. Manila, is forty-five dollars 
per ton. 

In this connection it seems to me that the question of the electrolytic decomposi- 
tion of salt brine for the simultaneous production of liquors for alkaline digestion 
and for bleaching should seriously be considered, for the local salt would thus be 
used for furnishing an expensive commodity otherwise imported, and the problem 
would be solved of manufacturing bleached fiber, which otherwise would be out of 
the question locally because of 'the extremely high freight rates on bleaching 
powder and its great tendency to lose strength during transit and storage. 

The commercial production of alkalies and bleach by this process, although of 
comparatively recent origin, may be considered as well past the experimental 
stage. The Townsend cell used at Niagara for the electrolytic production of alkali 
and chlorine has been in continuous operation for over four years, and a high 
efficiency and low initial and maintenance costs are claimed for it. The strength 
of the caustic liquor can be regulated at will and the cell is on the market in 
cslpacities of from 2,000 to 6,000 ampere units. 

The Whiting process for the electrolytic production of alkali and chlorine is 
regarded as exceedingly economical in its use of power, and to require a minimum 
amount of labor in its operation. It was recently developed by Jasper Whiting, 

''Approximately 1,200 tons of lime are annually marketed in Manila at the 
present time. This is for the greater part water-slaked lime made from coral 
deposits at Malabon on Manila Bay, and from the stone quarries at Binangonan 
on Bay Lake. A much better source of limestone of a good grade for the proposed 
industry is in the marble outcrops on Romblon Island, situated about 100 miles 
from Manila and in nearly a direct line with the local coal supply. The deposits 
are situated on a well-protected, deep-water harbor. The proximity of the town 
of Romblon insures labor for quarrying and hauling. The deposits are practically 
unlimited and the stone burns to a particularly fat lime, especially free from 
iron, sand, and magnesia. As there is no fuel at Romblon, the limestone would 
have to be transported to a more favorable place for burning. 
• The following information is given as a basis for estimating the cost of quarry- 
ing and transporting stone under local conditions: The city of Manila is paying 
one peso and twenty- five centavos per cubic meter of road material in quantities of 
50,000 cubic meters (approximately 75,000 tons) annually. This cost figure in- 
cludes quarrying, steam-crushing, transporting for about 37 kilometers by water 
and ulnoading. Nearly two-fifths of the above cost is for steam-crushing, which is 
unnecessary for limestone for burning. It is estimated that quicklime could be 
produced for about the maximum cost of manufacture elsewhere, viz, two pesos 
and ninety centavos per ton. These figures are based upon the cost of quarrying 
and transporting limestone from Romblon to the Bataan coast adjacent to mill 
site at one dollar, and upon the local cost of labor and fuel. 



PmLIPPINE FIBERS AND FIBROUS SUBSTANCES. 249 

of Boston, and installed in the Oxford Paper Company's mills, at Rumford Falls, 
Maine, and in the soda pulp .mill of the New York and Pennsylvania Company, at 
Johnsonsburg, New York. The latter company manufactures 75 tons of soda pulp 
per day, and makes all its alkali and bleach liquors by electrolysis. 

At the present time onl}'^ about 800 tons of crude sea salt are manufac- 
tured by solar evaporation of sea water on the phores of Manila bay. 
A 20-ton soda pulp mill would require approximately 4 tons of salt per 
day, hence the present small capacity for production would need to be 
considerably increased. 

BEPAIRS, RENEWALS, AND DEPEECIA%?ION. 

These items, the world over, are heavy in the pulp' industry. The 
average life of a soda pulp mill is variously estimated at from seven to 
ten years. The great distance from the base of supplies will necessitate 
the carrying in stock of a full line of spare parts which, of course, 
augments the monthly repair cost item by the amount of insurance and 
interest on the investment. 

TAXES, INSURANCE, AND INTEREST. 

Bamboo is not a taxed forest product at the present time. In the 
event of its commercial exploitation an internal-revenue tax of about 
10 per cent of the market value of the bamboo would probably be 
levied; this, based upon the present market price, would be, on the 
quantity required to produce a ton of pulp, 25 cents United States 
currency. 

At the present time an internal-revenue tax of 0.33 per cent is also charged on 
the gross sales of any local business. Data are not at hand for comparing the 
cost of property taxes, insurance, and interest with similar charges elsewhere. 
Manifestly, in calculating the total cost of production per ton of product, the 
annual taxes levied on the plant, insurance carried, and the interest on the 
investment should enter. 

It will be seen from the above analysis of the cost of manufacture, 
that any commercial advantage possessed by this locality as a suitable 
place for a pulp industry must lie in the other item which goes to make 
up the total cost of production, namely, that of the raw material. 

The following is an estimate, in dollars United States currency, of 
the cost of a mill designed to manufacture paper pulp from bamboo.^^ 

*^ This estimate, in so far as it regards the equipment, is based upon itemized 
estimates from several American and European sources, carefully revised by Her- 
bert S. Kimball, chemical engineer and mill architect, of Boston, Massachusetts. 
I wish to take this opportunity of heartily thanking him for his courtesy. 
96918 2 



250 RICHMOND. 

EQUIPMENT. 
I. IN THE MILL YARD. 

Donkey engine and hoisting crane 1,000 

Trucks and tracking 1,000 

II. BAMBOO-PREPARING PLANT (WOOD ROOM). 

Slasher 1,000 

Conveyor (to chipper) 100 

Chipper 1 : 750 

Crusher 600 

Chip screen ^ 800 

Conveyor to chip bin 500 

Chip bin 200 

Knife grinder 300 

III. ALKALI-MAKING PLANT. 

Battery of leaching pans for black ash 2,000 

Pump for leacher 110 

Tank for recovered soda solution 300 

Elevator and scales 400 

Six causticizing tanks with stirrers 3,000 

Pump for strong and weak liquors 110 

Two storage tanks for cooking liquors 1,400 

Pump to fill digester 200 

IV. SODA-RECOVERY PLANT. 

Storage tank for weak black liquor from digester 400 

Pump to evaporator supply tank and for wash pits 110 

Evaporator supply tank 300 

Pump to tank supplying furnace 110 

Multiple effect evaporator 12,000 

Tank supplying black ash furnace 200 

Black ash furnace 3,000 

Steam boiler to utilize waste heat from furnace 1,800 

Conveyor for black ash to leach tanks 200 

V. PULP-MAKING PLANT. 

Two digesters, 7 feet by 29 feet, at $4,370 8,740 

Blow pit 1,000 

Six wash tanks (steel) 2,400 

Wooden stuff chest (with stirrer) 1,000 

Pump for stock •- 200 

Riffler 200 

Screen : 650 

Washer 1,000 

Stuff chest for machine 500 

Pump for white paper 110 

Puipp for stock to machine 110 

Pulp-drying machine, complete 15,000 

Press (knuckle joint type) 300 

Scales 100 



PHIUPPINE FIBERS AND FIBROUS SUBSTANCES. 261 

VI. POWER PLANT. 

Steam engine for main mill, 325-horsepower 6,000 

Steam engine for wood -preparing room, lOO-horsepower 1,500 

Four steam boilers, each 200-horsepower 7,500 

Feed water heater .: ^ 

Feed pump I 3,000 

Other boiler-house equipment except piping ) 

Main pump _ 300 

Tank for fresh wq,ter 800 

Tank for weak soda washings : 600 

Shaftings, belting, etc. (all power transmission), with labor.. 4,000 

Piping through mill with labor 7,000 

Filter plant 2,600 

Electric lighting plant, wiring, etc 5,000 

VII. MISCELLANEOUS ITEMS. 

Foundations for machinery 1,000 

Labor not counted for, (erecting machine, etc.) , 1,000 

Fire protection: hydrants, pipe, and pump 2,600 - 

Shop and tools 1,500 

Spare parts and repair stock 1,000 

Office fittings 1,100 

Total : 110,000 

Cost of construction work, buildings, installation, etc., based 
on estimates on specifications submitted to local con- 
tractors 25,000 

Launch carrying supplies, towing coal and limestone barges.. 5,000 
Barges, limekiln, trucks and portable tracking from tide 

water to mill site 20,000 

Working capital 40,000 

Total 200,000 

CONCLUSIONS. 

It is believed that 200,000 dollars United States currency is sufficient to 
build, equip, and operate a soda pulp mill of 20 tons capaxjity daily. 
Such a plant built in a modern manner and efficiently operated will 
produce unbleached air-dry bamboo fiber at a maximum cost of 21 dollars 
per short ton, f. o. b. Manila. If the excellent quality of bamboo soda 
pulp^ prepared under favorable conditions is considered, a profitable export 
trade with Japan^ Europe, Australia, and the Pacific coast, in direct 
competition with chemical wood pulp at present quotations, appears a 
not unreasonable expectation. 

One consideration it seems to me is of especial importance, and that 
is the reasonable possibility of obtaining the raw material in such quan- 
tity and at such a figure as to allow of manufacture at a profit even 
imder conditions of maximum cost. 

Tlie above opinions are the result of several years of laboratory in- 
vestigations of the raw materials for pulp making, and the problems 



252 RICHMOND. 

connected therewith, supplemented by careful field suiTeys, and due 
regard for the local conditions bearing upon the subject. 

PHILIPPINE PAPEK-MAKINO MATERIALS. 

A revival of interest has recently been shown in cogon grass as a 
source for paper stock, by a request from a leading paper manufacturer 
in the United States for 2 or 3 tons of the baled material for a thorough 
trial on a commercial scale. Cogon grass, it will be remembered, was 
studied and reported upon in This Journal for June, 1906.^^ 

At the time when several large estates in the Malay Peninsula were 
cleared for the planting of rubber trees, attention was directed to the 
possible use of a- grass called lalang, indigenous to that locality. Doctor 
Foxworthy, botanist of the Bureau of Science, who recently has returned 
from an extended trip throughout the Malayan regions, makes the fol- 
lowing statement regarding it : "One of the few things of which we are 
absolutely certain is that lalang and Philippine cogon are botanically 
identical. ^^ 

Clayton Beadle, a British pulp and paper expert, published in the 
World^s Paper Trade Eeview for July, 1907, the following report on 
the use of lalang grass for paper, based upon an investigation made by 
'him in 1891: 

Experiments on 50 pounds of material showed that the grass could be cooked 
to a pulp in five hours at 50 pounds pressure with 15 per cent of its weight 
of caustic soda. It bleached white, consuming 10 per cent of its weight of 
bleaching powder. The fiber is longer and thinner than esparto grags fiber, 
bulks equally well and gives a tougher sheet. Unbleached, sized, and mixed 
papers of lalang grass pulp have not deteriorated since they were made, about 
fifteen years ago. The paper takes a good surface under the calendar, and will 
carry the usual amount of loading. The fiber yields 46 per cent cellulose as 
* compared with 48 per cent from esparto, and the cost of treatment is about 
the same." 

This report on the suitability of lalang grass for paper making was* 
further confirmed by Kemington, in 1908, whose investigation is given in 
full: 

Lalang {Imperata arundinacea Cyrill.) as a Paper-making Material. 

One of our correspondents sent a sample of lalang grass recently to England, 
and has received the following report, also samples of paper made from lalang 
grass entirely and one-half lalang and one-half cotton: • 

CERTIFICATE OF ANALYSIS. 

SAMPLE OP LALANG GRASS MARKED "EXCHANGE, SINGAPORE, STRAITS SETTLEMENT," 

ON 23D JULY, 1908. 

This is to certify that the above sample has been carefully examined with 
the following results : 

This grass was forwarded to *'The Aynsome Technical Laboratories" for in- 

''This Journal (1906), 1, 457. 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 253 

vestigation, that its commercial use as a source of pulp for paper making might 
be ascertained. 

The object of the present report is to draw attention to the importance and 
value of this grass as a pulp-producing material, and to indicate its chief 
characteristics and economic value. 

The sample was a pale buff color, lustrous in appearance and of fair strength, 
of the order Gramineae. 

The chemical examination furnished the following results, the percentages 
other than that of moisture being expressed on the dry material: 

Moisture 13.21 

Ash 4.14 

Loss on L. hydrolysis 10.76 

Loss on B. hydrolysis 46.65 

Loss on mercerisation 31.62 

Loss on acid purification 0.95 

Gain on nitration 21.19 

Cellulose 47.41 

Length of ultimate fiber (m'm.) 1.20 

The ultimate fiber obtained from this grass is very similar in most respects 
to esparto; the yield of bleached fiber being about the same. This is a favorable 
indication, inasmuch as "esparto" is one of the best known and most useful 
sources of supply to the trade. The fibers as seen under the microscope are 
short, smooth, cylindrical, fairly uniform in diameter, gradually tapering to 
rounded extremities; they also occur together in little bundles. The pulp will 
be found to contain a number of small cuticular cells which, however, do not 
show in the finished paper. The fibers are stained a pale yellow with iodine 
solution, which fades more rapidly than is usual with colored pulps. 

The results obtained from the chemical analysis show that the grass is capable 
of yielding a good quality of cellulose, suitable in every way for the manufac- 
ture of paper. Although the grass is very susceptible to the action of dilute 
alkalis, the final product is exceptionally pure and readily resolved. 

From observations noted during this preliminary examination, the following 
scheme was adopted for the production of the pulp on a larger scale. 

PROCESS. 

The available grass in its natural condition weighing 400 grams (14.1 ounces) 
was in a clean state, and required little treatment beyond cutting into small 
pieces ready for boiling. It is usually necessary on a large scale to pass the 
material through some type of cleaner to remove dirt and adventitious matter. 
The grass was then thoroughly wetted and soaked until it became soft and 
pliable. It was placed in a boiler of the spherical type, covered with water and 
digested with caustic soda corresponding to 15 per ceiit on the grass treated 
for a period of ten hours under a pressure which was kept constant at 4 atmos- 
pheres; an even temperature about 135° C. being maintained. 

Owing to thfe quantity of material at our disposal being somewhat limited, 
it was only possible to conduct one experiment,- consequently a good margin 
in the use of caustic and degree of general treatment was allowed. 

The pulp obtained after washing was of good uniform quality and color. The 
yield agreed very closely and was only slightly higher than the preliminary 
chemical analysis showed, which indicates complete reaction in the digester. 

The pulp was carefully beaten for about an hour and at the same time bleached, 
chloride of lime being used for the purpose: 10 parts of dry powder being 
used per 100 parts of pulp. 



254 



RICHMOND. 



The stuff was taken from the beater and well washed and subsequently a 
small quantity of loading was gradually added and the whole again beaten 
for half an hour. At this stage of the operation the rosin size was introduced 
and the decomposition of the soda resinate completed with the calculated quan- 
tity of alum. The amount of size used corresponded to 4 per cent on the 
dry pulp present, the total time occupied in preparation of the pulp for running 
on the machine being two hours. 

PBEPABATION OF PAPER. 

No great difficulty was experienced in running the pulp; it retained, however, • 
a considerable quantity of water after passing the suction boxes and in conse- 
quence it was found necessary to keep the press roll down hard. On a large 
machine this precaution would not be necessary, as more suction boxes are 
available and complete control is assured. 

The paper was passed over 9 cylinders at a pressure of about 8 pounds to the 
square inch, then through one calendar and finally reeled off. In a similar 
manner a second sample was prepared, using a mixture of half pulp obtained 
from the lalang grass and half cotton beaten together. The paper obtained from 
this blend could prove very useful as a high-class wrapping paper, it being 
stronger and possessing a comparatively high resistance to folding, as will be . 
seen in the following table of physical tests done on the finished specimens of 
papers. The paper made from "all-grass" pulp would with judicious treatment 
for improvement of color be very suitable for printing purposes. 

The following table of figures was obtained and compiled from tests and 
analyses made on the air-dried finished papers. 



Pure 
lalang. 



Half 

lalang, 

Half cotton. 



Physical properties: 

Breaking strain pounds— 

Stretching ^ per cent- 
Rubbing test percent__ 

Thlcltness inches__ 

Chemical constituents: 

Rosin size l per cent__ 

Ash (natural and loading) per cent- 
Moisture ^- percent-- 



8.37 

1.03 

.27 

1/125 

^2.20 

b3.01 

b 10. 40 



•8.88 
• 1.35 
.151 
1/234 

b2.43 
bl.94 
b8.68 



a Way of machine. 



^ On the dry paper. 



The addition of cotton to the pure lalang fiber has the effect of improving its 
resistance to crumbling, and also in a lesser degree its tensile strength. 

(Signed) G. Stewabt Remington. 

It would seem that these reports are final and conclusive in regard 
to the suitability of this material for the purpose in question, and that 
further experimental work is entirely unnecessary. We find the grass 
growing wild in the Philippines and in such abundance in certain local- 
ities as to warrant a serious inquiry looking to its. commercial use. 
Preliminary surveys point to the Island of Masbate, and possibly the 
neighboring Islands of Ticao and Burias, as the most favorable loca- 



V. 



PHILIPPINE FIBERS AND FIBROUS SUBSTANCES. 255 

tions for a paper-pulp industry leaking use of this grass, for the follow- 
ing reasons : 

First. Masbate possesses extensive areas of rolling lands covered with 
comparatively clear stands of cogon grass. 

Second. There are numerous well-protected deep-water harbors ad- 
jacent to the grass lands. 

Third. Clear water for steaming and manufacturing purposes is avail- 
able. 

Fourth. Large undeveloped deposits of a good. grade of limestone are 
at hand in several places. 

Fifth. Coal of fair quality has been located near Cataingan on the 
east coast and mangrove wood for fuel is abundant in or near all of 
the harbors. 

In the event of serious inquiry by prospective investors, the following 
points in Masbate should receive attention in the order named: Aroroy, 
on the north, Mandawan on the west, and Cataingan on the east coast. 
The latter has the chief advantage in respect to the local coal deposits, 
but is at a disadvantage because of large private land grants employed 
for grazing purposes at the present time. 

Labor is scarcer in Masbate than in the majority of the other islands 
and offers the only serious difficulty which would be encountered. 



PHILIPPINE TERPENES AND ESSENTIAL OILS, IV. 



By Raymond F. Bacon. 
(From the Chemical Laboratory, Bureau of Science, Manila, P. I.) 



The following notes on various Philippine oils represent such data 
as have accumulated since my last publication on this subject.^ 

OIL OF CINNAMOMUM MINDANAENSE ELMER. 

A certain quantity of the bark of this tree was collected by Mr. Elmer 
in the hills behind Davao^ Mindanao. It is fairly common in that region. 
The tree is very close to Cinnamomum zeylanicum Nees and the bark 
in appearance, taste, and odor is just lil^e the cinnamon of commerce. 
Fifty kilos of the ground bark were distilled with steam, yielding 200 
grams of oil of a yellow color and of a strong cinnamon odor and taste. 

This probably does not represent by any means all of the oil which 

it is possible to obtain by commercial distillation from this bark, the 

proportion being less because of the small amount of material at' 

my disposal. The oil had the following properties: Eefractive index, 

30° 30° 30° 

N -jT- 1.5300; optical rotation, A -^- If.^d; specific gravity, -^^ 0.960. 

Ten grams of the oil gave 9.2 grams of the dry sodium bisulphite 
compound of cinnamic aldehyde, corresponding to an aldehyde content 
of approximately 60 per cent. 

This oil does not agree very closely in its physical properties with 
the Ceylon cinnamon oil from C. zeylanicum. 

OIL FROM CANARIUM VILLOSUM F. VILL. 

In a previous publication ^ on this subject I mentioned the oil from 
the resin of this tree, known locally as pagsainguin, and stated that it 
consisted principally of paracymol; this has also been confirmed by 
Schimmel and Company on a sample sent to them by this laboratory. 

In November, 1909, 3.5 kilos of resin were collected from one tree 
near Lamao, Bataan Province. The volatile , oil was distilled from ihe 

^This Journal, Sec. A (1909), 4, 93. 
^This Journal, Sec. A (1909), 4, 94. 

267 



258 



BACON. 



resin in vacuo (4 to 6 millimeters) giving a total of 390 grams of oil 
(about 11 per cent). The latter had an odor like that obtained duiing 
a similar distillation of Manila elemi (C^ luzonicum A. Gray)^ and the 
aqueous portion of the distillate contained a considerable amount of 
formic acid^ although there were no visible evidences of decomposition 
of the resin during the distillation in vacuo. The oil was then distilled 
six times over sodium, using a column of glass beads in the neck of the 
distilling flask, and gave the following fractions: 



Fraction No. 


Weight. 


Boiling 
point. 


Refractive 

30 

index, N-fs- 


Specific 

.. 30 
gravity, g^- 


Optical ro- 

30 

tation, A j). 


1 __ . ^_ 


Grams. 
102 
19 
45 
40 
47 
10 


o 
154-158 
158-161 
161-165 
165-170 
170-175 
175-180 


1. 4645 
1.4660 
1. 4690 
1.4730 
1.4770 
1. 4795 


852 .S9.4 


2 


851 
850 
849 
819 


34.7 


3 - 


29.4 


4 ,___ 


21.1 
13.5 


5 

6 







The residue was a thick, brown, viscous oil, which was attacked by 
sodium when an attempt was made to distill it over that metal. 

Fraction No. 1 had a strong odor of pinene, and Nos. 1, 2, and 3 each 
readily gave large yields of pinene hydrochloride, melting at 125°. 
From the higher boiling fractions a very small quantity of dipeniene 
was obtained, the tetrabromide melting at 124°. 

The principal constituent of the volatile oil of this sample of the 
pagsainguin resin is therefore c?-pinene. In a previous paper ^ I have 
shown, from the examination of a very large number of specimens of 
Manila elemi from individual trees, that the terpenes found in these 
trees vary quite markedly from tree to tree, and that one tree usually 
yields but a single terpene. The same would probably seem to hold 
good for the pagsainguin resin, and it appears probable that these 
Ganarium trees manufacture a large series of terpenes and also the parent 
substance of terpenes, p-cymol. The next step will be to study the resin 
from one tree for a considerable length of time, to discover whether, 
for example, an individual tree always gives a resin containing pinene, 
or whether at one time it yields a product having pinene as a constituent, 
at another, a resin containing phellandrene, etc. This work will be 
continued in this laboratory, as the resin should be of considerable value 
in making clear the physiologic process of the plant in the formation of 
resins. 



^ Loc. cit. 



PHILIPPINE TERPENES AND ESSENTIAL OILS, IV, 259 

OIL OF ZINGIBER OFFICINALE ROS€. 

The ginger plant is found in aJl parts of the Philippines^ being culti- 
vated by the natives in small patches for use as a condiment for food. 
It is not grown on a large scale and, so far as I can discover, ginger has 
never been exported. I made two experiments on the distillation of 
native ginger roots. In the first one, 50 kilos of the chopped roots 
gave only 25 grams of oil. For the second, 132 kilos were purchased in 
the market at Malabon at 22 centavos, (11 cents United IStates currency) 
per kilo and immediately distilled. There were obtmned 95* grams (0.072 
per cent) of a light yellow oil, having the odor of ginger and also a 
strong smell, much like that of orange-peel oil. This oil had the follow- 

qr\o 30° 

ing properties : Specific gravity ^ 0.8850 ; refractive index, JST -g- 

30° 
1.4830; optical rotation, A -jy 5.°9; saponification number, 14;. 

It is easily and completely soluble in two or more volumes of 90 per 
cent alcohol. 

It is seen that oil from the Philippine ginger differs quite markedly 
in its properties from that distilled from the Jamaica or African varieties 
and resembles some Japanese oils examined by Schimmel and Company * 
in its ready solubility in 90 per cent alcohol, and its negative optical 
rotation. 

As there have been many inquiries at the laboratory with regard to 
the planting of ginger and its utilization, I also present the following 
data on this subject. 

Zimmermann '' publishes in classified form the results qf the observations and 
experience up to .the present day on the cultivation and preparation of ginger. 
The mother-plant of ginger, which is indigenous to tropical Asia and is cultivated 
in numerous countries of the Tropics, is, as is well known, Zingiber officinale 
Rose. The rhizomes of this plant contain, according to their derivation, different 
quantities of essential oil ; whereas African ginger yields 2 to 3 per cent, Jamaica 
ginger only yields 1.075 per cent essential oil. With regard to the fertility of 
the soil, the requirements of ginger are fairly high: it must not be too firm, and 
not swampy. A sandy loam .which is also chalky is the most favorable for its 
cultivation. Ginger is grown exclusively from pieces of rhizomes, which are 
kept in dry places und which, shortly before sowing, are cut up in bits from 2.8 
to 5 centimeters in length, and each piece must have at least one bud. The fields 
are laid out in .the same manner as a potato field, most suitably with ridges 
of 30 centimeters and furrows of about 68 centimeters in width. The pieces of 

* Semi-annual Report. October (1893), 46; Gildemeister & Hoffman. Die 
Aetherischen Oele (1899), 406. 

'^ Communications from the Biologico- Agricultural Institute, Amani. Reprint 
from the Usumhara Post (1904), No. 28. 



260 BACON. 

tuber are placed on the ridges in holes about 7 to 10 centimeters deep, and 25 
to 30 centimeters apart, which are well filled up with earth so that the tubers 
lying in the holes do not decay. 

The harvest commences when the parts which are above the ground are 
withering, which is usually the case after nine to eleven months; the tubers 
are then removed from the soil. The further treatment always begins with 
cutting off the roots from the carefully washed tubers. From that point the 
treatment varies, according as it is desired to prepare either dried or preserved 
ginger. In the dried ginger a distinction is again made between peeled or white, 
and unpeeled or black ginger. 

In preparing peeled ginger, only the thinnest possible skin may be removed, 
as the aromatic constituents are present in the portion close below the epidermis. 
After peeling, the tubers are immediately placed again in clean water in which 
they are left overnight, and are then again dried. As in the case of peeled 
ginger, great value is attached to a bright and, as much as possible to a white 
color. Chemical agents (chloride of lime, plaster of Paris) have been tried to 
improve the color of the ginger tubers but it is strongly advised not to do this. 

The unpeeled ginger, after having been carefully cleaned, is immediately dried. 

The preparation of preserved ginger takes place as follows: The washed, boiled, 
and peeled tubers are placed in earthenware vessels, and a boiling solution of 
sugar (1 kilo sugar to 2 liters water) is immediately poured over them. After 
twenty-four hours this solution is drained off, heated and poured back into the 
vessel. This is repeated after two days. The ginger can then be taken out of 
the solution of sugar, and dried or sent away directly in jars. 

The yield from a ginger field amounts to about 1,100 to 1,700 kilos per 
hectare, but it is said that, exceptionally, 2,200 kilos per hectare have been 
gathered. At the price of ginger a considerable profit per acre might be expected, 
but the cultivation requires much manual labor which makes it too expensive to 
grow ginger on plantations; moreover, the ginger plants very greatly exhaust 
the soil, so that repeated cultivation on the same soil is only possible with a 
considerable amount of fertilization. 

The dried root is valued at 20 to 30 shillings per hundredweight on the London 
market, and London broker^ say that ginger root must be quite dry in order to 
be saleable on that market. 

Spice ginger is the ground rootstock. Preserved ginger root is prepared by 
boiling the roots in water and curing with sugar. Much of this commodity 
comes from Canton. Extract of Jamaica ginger is made by extracting the ground 
roots with strong alcohol. The residue left in the production of the extract of 
Jamaica ginger or of ginger ale is termed exhausted ginger and is used as an 
adulterant in ground spice. 

It appears that certain, kinds of ginger are more suitable for distilla- 
tion than- others. Schimmel and Company distill only the African variety, 
imported from Liberia, as it contains from two to three' times as much 
oil as Jamaica ginger. 

Eecent quotations in the United States are : Jamaica unbleached ginger 
root, 12 to 15 cents per pound; Jamaica bleached, 16 to 17. 

Under the Payne Bill tariff, unmanufactured ginger is placed on the 
free list. 



PHILIPPINE TERPENES AND ESSENTIAL OILS, IV. 261 

OIL OF OCIMUM SANCTUM L. 

This plant is found cultivated in many of the gardens of the Philip- 
pines. From 13.86 kilos of leaves, which were forty-eight hours old at 
the time of distillation, 83.3 grams of a green-colored oil (0.6 per cent) 
were obtained. This oil had a sweet, anise-like odor and the following 

qno 30^ 

properties: Eefractive index, N -jr- 1.5070; optical rotation, A -jt- 0; 

specific gravity, ^^ 0.952; saponification number, 2.8. 

. The oil consists to a large extent of methyl homoanisic acid, melting 
at 85°, being obtained by oxidizing the fraction boiling from 85° to 95° 
at 9 millimeters pressure. I will state concerning the amount of oil which 
it is possible to obtain frOm a given number of these plants, that 65 small 
plants in the yard of the Bureau of Science gave 2.5 kilos fresh leaves 
which yielded, on steam distillation, 32 grams of oil. . This represented 
about 60 per cent of the total leaves of these plants, which were about 
two years old, but they had been transplanted several times during those 
two years. 

OIL OF CURCUMA ZEDOARIA ROSC. 

This plant is quite abundant in many places near Manila and is found 
pretty well distributed over the Philippines. I do not know that it is 
used for any purpose by the Filipinos. 

One hundred kilos of chopped roots distilled with steam gave 65 grams 
(0.065 per cent) of a brown oil, which was too dark accurately to 
determine its optical rotation. Its other properties were as follows: 

QAO 30° 

Specific gravity, -^ 0.933; refractive index, N -^- 1.4920. 

The oil is soluble in two or more volumes of 90 per cent alcohol. 
These roots were obtained in September. On the 1st of December, 160 
kilos of roots were obtained, chopped up and distilled with steam. 
They gave 400 grams (0.25 per cent) of an oil of a light yellow color, 

30° 

and having the following .properties: Specific gravity, —j- 0.993; re- 

OAO gQO 

fractive index, N ^r 1.5070; optical rotation, A -g- 1°.10; saponification 

number, 2, soluble in two or more volumes 80 per cent alcohol. Two 
hundred and fifty grams of this oil were distilled in vacuo (7 milli- 
meters) and gave the following fractions. 



Fraction. 


BoiUng point. 


Weight 
(grams.) 


1 


60-100 


42 


2 


100-120 


38 


3 


120-140 


16 


4 


140-160 


27 


5 


160-166 


104 



262 BACON. 

A solid, crystalline residue weighing 20 grams was left in the distil- 
ling flask. 

Fractions 4 and 5 solidified to give a beautiful, white, crystalinne 
compound* which most probably is a sesquiterpene alcohol. 

This alcohol is exceedingly soluble in all organic solvents, being recry stall ized 
from low-boiling petroleum ether. The crystals from alcohol are very frequently 
several centimeters in length. The body melts sharply at 67" and boils at 
about 160" (7 millimeters), the exact boiling point being obscure because the 
substance sublimes at a temperature quite a number of degrees below the latter. 
The crystalline body possesses a strong and rather pleasant odor, and is undoubt- 
edly the substance which gives its characteristic smell to the * oil of zeodoaria 
Some of the crystals which were left in a tightly stoppered test tube for three 
months partially liquefied and the characteristic odor had then been replaced 
by one resembling turpentine. The sesquiterpene alcohol when dissolved in 

alcohol or benzene shows no optical activity; its specific gravity, _—, is 1.01 so 

that it probably belongs to the tricyclic system of sesquiterpene alcohols. With 
concentrated sulphuric acid at first it gives a deep red color and then chars; 
a white, solid, odorless substance is formed by hot,' concentrated nitric acid. It 
was not possible to obtain this substance in a crystalline state, it dissolves in 
10 per cent sodium hydroxide to give a red solution and is precipitated by acids 
as a white, amorphous solid. Phosphorus pentoxide was added to a solution 
of the sesquiterpene alcohol in petroleum ether without any apparent action; 
following this I added one drop of strong formic acid, whereupon a hard resin, 
in appearance and general properties much like the varnish resins, resulted. As 
this beautifully crystalline sesquiterpene alcohol is available in large quantities, 
its study will be continued. 

OIL OF. CURCUMA LONGA L. 

One hundred and twenty-three kilos of Philippine roots from this 

plant were distilled with constant cohobation, yielding 290 grams of a 

brown colored oil with the followiiig properties: Specific gravity, 

30® 30° S0° 

g^-o 0.930; refractive index, N -^ 1.5030; optical rotation, A -jy8°.6; 

ester number, 81. 

Miscible with 75 per cent or stronger alcohol in any proportion.^ 

OIL or MICHELIA CHAMPACA L. 

The small amount of literature on this subject is of a very confusing 
^nature, and I believe that the oils which have been put on the market 
as champax^a oils have their origin only to a very slight extent in Michelia 
champaca. 

In 1882, Schimmel and Company^ report a champaca oil which was liquid 

•Jackson, Amer. Ckem. Journ. (1882), 4, 368, gives the following numbers for 
curcuma oil; Ao— 24.68" jd=0.95.61. 

Rupe, Lusch and Steinbach, Ber, d. chem. Ges. (1909), 42, 2575, give a specific 
gravity at 20° of 0.9388; Ajj, — 24.70. The Philippine oil seems to be somewhat 
low in optical rotation as compared to the oils examined by these investigators. 

^Semiannual Report (1882), 1,7. 



PHILIPPINE TERPENES AND ESSENTIAL OILS, IV. 263 

v/hile warm, but when cooled, solidified to a crystalline mass, like rose oiL In 
1894 ^ the same firm speak of a champaca oil of a reddish-yellow color, with an odor 
reminding one somewhat of cassia flowers, which also separated considerable 
quantities of crystals. In 1897* this firm obtained a champaca oil originating 
in Manila, of specific gravity of 0.938, optical rotation, — 68°.2 and ester number 
77.3. They also state that the oil had a slight resemblance in odor to that of 
ylang-ylang and, like the latter, contained benzoic acid. In 1907 *° the same 
firm report the following substances as occurring in champaca oil: Linalo5l, about 
60 per cent; small amounts of geraniol and of eugenol methyl ether; and in the 
lower boiling portions, the methyl ester of methyl ethyl acetic acid. 

In so far as. Manila oils are cbncemed, I consider that these reports 
do not represent an oil from the yellow champaca flower, as so far as I am 
able to learn there never has been a time in Manila when as much as 1,000 
kilos of flowers were available at any one time. - 

I have worked only with the oil from the yellow champaca, and have 
not made any studies on the one from the white flowers. It is probable 
that these reported oils represent distillations of champaca mixed with 
other flowers. In the Manila oil of 1897 these other flowers were prob- 
ably ylang-ylang, because of the odor and physical properties of the 
oil and of the presence of benzoic acid, which I do not find in my cham» 
paca oils. 

Several synthetic oils have been announced, the only one of which I 
have seen possessed a very poor imitation of the true champaca odor. 
Our investigations on this oil have not advanced very fai;, but some data 
which may be of interest will be recorded. 

The yield appears to be over 0.2 per cent. The crude oil on standing separated 
a large amount of a crystalline solid. This was filtered and an additional 
qiiantity of it was again separated by the addition of ether, in which the solid 
is quite insoluble. The remaining oil, after standing for some weeks in the 
laboratory, continued to solidify until it gradually became semisolid. The second ' 
solid which separated was amorphous and appeared to be resinous in nature. If 
this semisolid extract is treated with 70 per cent alcohol, about half of it separates 
in the form of the amorphous, brown, odorless body. This was filtered 'and the 
filtrate concentrated at 40° in vacuo until a brown oil separated, which had a 
very fine odor of champaca, and was readily soluble in 70 per cent alcohol or 

stronger. Our oil had the following constants: Specific gravity, pL, 0.9643; 

Of\0 

refractive index, N _-_, 1.4550; saponification number, 100. Another oil had 

qno or»o 

specific gravity, _ -, 1.020; refractive index, N _-, 1.4830; saponification number, 

180. The second oil had the finer odor. The oils were too dark to permit of 
determinations of the optical activity. My champaca oil is neutral. The recent 
oil examined by Schimmel and Company" contained a considerable proportion of 
acid, especially of methyl ethyl acetic acid. 

'^Ibid. (1894), 1, 58. 
^Ibid. (1897), 1, 11. 
,^o/6id. (1907), 2. 
" Loc. cit. 



264 BACON. 

Fifty grams of champaca oil (soluble in 70 per cent alcohol, ester number 
180) were saponified with 10 grams potassium hydrate in 100 cubic centimeters 
of 96 per cent alcohol. After heating for one hour with a reflux condenser, 
two volumes of water were added; 4.5 grams of an amorphous solid separated. 
This was filtered and the filtrate was separate into neutral, acid, and phenol 
fractions. By saponification the champaca oil loses all of its characteristic odor, 
which therefore must be due to esters. 

The phenol fraction (1.5 grams) proved to consist principally of iso-eugenol, 
as benzoyl iso-eugenol melting at 103° could be obtained from it. The total 
acid fraction weighed 15 grams. None of this acid boils below 140° at 40 milli- 
meters; hence there is no methyl ethyl acetic acid. No acids have as yet been 
identified. 

The neutral portion weighed 23 grams and had an odor somewhat similar to 
that of oil of bay. 

Some studies were made on the solid which crystallizes from the 
freshly prepared oil. This substance, as firs-t separated from champaca 
oil, is of a greenish color and possesses a faint, but pleasant odor. It 
is readily soluble in chloroform, acetone, acetic acid, hot alcohol or 
benzene. It is almost insoluble in ether or petroleum ether. It was 
■purified by repeated solution in hot benzene or chloroform and precipita- 
tion with ether or petroleum ether. A very peculiar phenomenon was 
noted during the purification. Many times, without apparent reason, 
the solution in benzene or chloroform would, with a loud, crackling noise, 
solidify to a horn-like jelly which was proved to be insoluble in all 
solvents, and it was impossible to regenerate the original body from 
this substance. Considerable quantities of the solid were lost in this 
manner and no method of preventing this polymerization (?) was 
discovered.^^ 

The crystalline solid, after repeated crystallization from benzene and 
petroleum ether, forms odorless, white needles. 

Eight grams were dissolved in a benzene-alcohol mixture and shaken with a 
concentrated 'solution of sodium bisulphite. A precipitate was formed which was 
filtered. The filtrate was evaporated to crystallization and 4.5 grams of a body 
obtained, which, after recrystallization from chloroform and ether, melted at 
165° to 166°. Combustion of this substance gave the following results: 

I. 0.2496 g. gave 0.1587 g. H^O and 0.5985 g. CO^. 

II. 0.2486 g. gave 0.1593 g. H^O and 0.5990 g. COo. 

Calculated for Found 

CiflHgoOs. I II 

per cent. per cent. per cent. 

0=: 65.7 65.70 65.40 

H= 6.8 7.12 6.80 

Molecular weight 292 , 

" Doctor Freer informs me that in working with acrylic acid several years 
ago he had a similar experience. 



PHILIPPINE TERPENES AND ESSENTIAL OILS, IV. 265 

A determination of the molecular weight was made by the cryoscopic 
method. Solvent 12.67 grams benzene. 





Freezing 
point. 


Substance 
(grams). 


Molecular 
weight 




3.790 












I 


3.706 


0.0692 


296 


II 


3.600 


0.1563 


298 



A semicabazone, crystallizing in white needles and melting at 205° to 206** 
with decomposition, and a phenylhydrazone, which separates in white leaflets, 
in appearance much like benzhydrol and which melts with decomposition at 153° 
to 155°, were obtained from the original crystalline substance. The phenylhy- 
drazone is only very moderately soluble in ether or petroleum ether, hot or 
cold; it is soluble in boiling benzene, nearly insoluble in cold; readily soluble in 
hot or cold ialcohol, chloroform, or acetone. 

The original ketone or aldehyde from champaca, with concentrated sulphuric 
acid gives a deep red colored substance, soluble in water to give a red solution. The 
determination of methoxy groups by ZeisePs method gave a negative result. The 
saponification of the substance with alcoholic potassium hydroxide showed that 
if the formula of Cir.HooOs is assumed, then two molecules of potassium hydroxide 
are used by one molecule of the substance. However, it does not appear as if 
this substance were an ester, because it gives several products on treatment 
with alcoholic potash, of which neutral substances form only a very small 
proportion. The body completely dissolves in alcoholic sulphuric acid with a 
red color and with the odor of ethyl acetate. Acetic acid is one of the products 
when it is oxidized with dilute potassium bichromate and sulphuric acid. 

The stndy of this substance is being continued. 

OIL OF YLANG-YLANG. 

A number of oils of ylang-ylang have been analyzed during the past 
year^ and the results prove my statement of three years ago still to hold 
good, that the quality of ylang-ylang oil can be judged with fair ac- 
curacy by a few simple analytical constants. The Manila oils which 
I examined during the past year were rather better than in previous 
years, due to strong competition of Keunion oils, to better distillation 
methods in the Philippines, and to a low price for the flowers. The 
ylang-ylang flowers sold as 7 centavos (3.5 cents United States currency) 
per kilo during this past season, as compared to about 40 to 50 centavos 
(20 to 25 cents United States currency) three years ago. This per- 
mitted the use during the past season of better flowers for distillation, 
and oils having ester numbers of from 130 to 150 were common. 

96918 3 



A PRELIMINARY STUDY OF THE EFFECT OF TROPICAL 

SUNLIGHT ON THE ATMOSPHERE, WITH SOME 

NOTES ON RADIOACTIVE PHENOMENA 

IN THE PHILIPPINES.' 



By Eaymond F. Bacon. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. /.) 



Early in 1906 I stated ^ that the tropical sunlight seemed to ionize 
the air so that it would discharge an electroscope at a more rapid rate 
than noninsolated air. Eecently, in connection with the general study of 
the effects of tropical sunlight undertaken in this Bureau, I have again 
taken up this question and confirmed my earlier results. 

The apparatus used was a Fontaktoscope according to C. Engler and H. 
)Sieveking, the instrument being accompanied by a table of voltages corresponding 
to the various positions of the aluminium leaves. With the diffusion cylinder 
hanging in the 10-liter can and attached to the rod carrying the aluminium 
leaves, the normal rate of fall of the electroscope corresponded to a loss of 
potential of 4.3 volts per hour, and, under the conditions given, this was fairly 
constant for a large number of measurements extending over about four mpnths, 
the extreme values being 4 and 5.5 volts per hour. • 

To measure the effect of light on the ionization of the air, the latter was 
drawn at a definite rate through the can in which the diffusion cylinder was 
suspended, the amount of air being measured by a gas meter. | 

Comparative measurements of the rate of fall of the aluminium leaves 
were then made at night in the laboratory, in the daytime in .the diffuse 
light of the laboratory and in the sunlight. When the apparatus was 
placed in the sunlight^ the electroscope proper was always covered, 
although this seemed to make practically no difference. Even if the 
apparatus were uncovered, no direct sunlight could penetrate to the 

^ Owing to the fact that Doctor Bacon has left the Philippines, the data so far 
obtained by him are now published, although the work as yet is very incomplete 
and is being continued by the Bureau of Science. (P. C. F.) 

*Thi8 Journal (1906), 1, 433. 

267 



268 



BACON. 



inside, as it is made of metal and ground glass. The results obtained 
are as follows: 

I. — In the dark. 



Date. 

February 21 

February 22 

February 24 : 

March 4 

March 7 



Time. 


Poten- 
tial, elec- 
troscope. 


p. m. 




9.45 


141.3 


10.15 


138.3 


9.30 


205.4 


10.30 


199.5 


9.18 


218.9 


10.18 


212.3 


9.00 


208.8 


10. 15 


199.5 


9.30 


215.6 


10. ,30 


209. 9 



Fall, 

volts per 

hour. 



6.0 
5.9 
6.6 
7.4 

5.7 



Air cur- 
rent per 
hour. 



Liters. 
1,000 

1,000 

1,000 

1,000 

1,000 



The average rate of fall for darkness, therefore, for a gas current of 1,000 
liters per hour, was 6.3 volts. 

II. — In the diffuse light of the laboratory. 



February 8 . 



February 9 _ 



February 12 _ 



p. m. 

1.58 

2. 58 

a. m. 

8.15 

8.45 

8.45 

9.15 

10.16 

10.46 

10.56 

11.16 



Poten- 
tial, elec- 
troscope. 



161.1 
147.9 

162.9 
154.8 
154.8 
146. 5 
149.9 
144.0 
141.9 
135. 9 



Fall, 

volts per 

hour. 



Air 

current 

per hour. 



14.2 
16.6 
11.8 
18.0 



Liters. 
900 

1,000 

1,000 

170 

1,200 



I 



Average for diffuse light, 15.6 volts. 



EFFECT OF TROPICAL SUNLIGHT ON THE ATMOSPHERE. 269 

III. — In sunlight. 



February 5 (»)_ 
February 5 (*')_ 

February 14 (<=) 
February 14 (<>) 

February 14 ___ 



Time. 
a. m. 


Poten- 
tial, elec- 
troscope. 




r 9.21 
\ 10. 01 


116.7 


90.5 


f 10. 10 
I 11. 10 


128.1 


82.3 


[ 9.15 


177.1 


\ 9.30 


167.1 


9.45 


161.1 


10.10 


159.9 


I 10. 25 


154.9 


10.40 


146.9 


10.57 


158.2 


11. 27 


142.5 


11.55 


134.1 


1 12. 55 


98.2 


1.20 


169.9 


1.50 


148.4 



Fail, 

volts per 

hour. 



37.1 
45.1 



Air 
current 
per hour. 



\ 20. 
} 32. 



31.^ 
35.^ 
43. ( 



Liters. 
1,000 

1,000 
640 

1,000 
1,000 

1,000 
1,000 
1,000 



» The sun was behind a cloud for five minutes during these measurements. 

^ During the first ten minutes of this measurement the sun was wholly unclouded and 
very bright, and during that time the electroscope fell at the rate of 76.2 volts per hour. 
During the remainder of the time the sun was intermittently covered with clouds. 

c The sun was behind clouds about one-quarter of the time. During the last fifteen 
minutes it was behind clouds aoout one-half of the time, 

d During the whole of the first fifteen minutes of this measurement the sun was behind a 
cloud of such density that the disc of the sun was just visible. During the second fifteen 
minutes the sun was behind dense clouds about half the time, and the other half it shone 
from a clear sky. The effect caused by the sun going behind a cloud was easily discernible 
in the instrument. 

Average for sunlight, 35.7 volts. 

The results seem to indicate with considerable clearness that the 
tropical sunlight ionizes the air. I am well aware that much more 
work must be done before a positive statement can be made on this sub- 
ject, especially as such phenomena do not seem to take place in temperate 
zones, although it is well known from the work of Seiiard ^ and others 
that certain forms of ultra-violet light do ionize gases to a niarked degree, 
even though Buisson/ working at an earlier date, was unable to detect 
any conductivity in air through which ultra-violet light was passing. 
Buisson's original paper is not available to me, so that I do not know 
the exact conditions under which he worked, and therefore am not able 
to discuss his results which seem to contradict those obtained by me. 
However, I incline to the view that the ionization observed may not be 
a direct effect of the tropical sunlight on the air, but may be a secondary 
phenomenon connected with the large ainount of moisture in the air of 
tropical islands and with the luxuriant vegetation of this region. Thus, 

'Drudes Ann. (1898), 1, 483; Ibid. (1900), 3, 298. 

* Quoted by Penn, in Ann. chim. et phys., Paris (1897), (7), 11, 526. 



270 • . . . BACON. 

it is well known that air is ionized by ' bubbling through water, the 
evaporation of water, and the formation of fresh-water surfaces being 
accompanied by the liberation of ions. 

C. T. R. Wilson ^ has studied the action of ultra-violet light on gases 
from the point of view of the effect produced by the light on the for- 
mation of clouds. The following discussion of his work by J. J. Thom- 
son ® would seem to be pertinent : 

Wilson found that with very feeble ultra-violet light, clouds were produced 
by expansion when this exceeded a certain definite amount, just as in the case 
of a gas ionized by Rontgen rays, and that the amount of expansion required was 
just the same for the ultra-violet light as for these rayS; this at first sight would 
make it appear as if the ultra-violet light ionized the gas. However, Wilson 
found that the clouds produced by ultra-violet light differed from those produced 
by Rontgen rays, inasmuch as the former were not affected by strong electric 
fields, whereas the formation of the latter was almost entirely prevented by such 
fields. H the clouds due to ultra-violet light had been due to the ionization of 
the gas, the ions would have been removed by the field and the clouds stopped. 
At the same time the coincidence between the expansions required for the forma- 
tion of clouds under ultra-violet light and the conditions when ions are present 
is so remarkable that we are very reluctant to believe that the nuclei are different 
in the two cases; it seems to me that an explanation in harmony with the facts 
is that charged ions do form the nuclei of the drops formed by weak ultra-violet 
light, but that these ions are produced during the expansion of the gas and are 
not present when the gas is at rest. These ions might arise in the following 
way: We have seen in Chapter VII that under the action of strong ultra-violet 
light visible clouds are formed without expansion, these clouds probably being 
due to the formation of hydrogen peroxide, which, mixing with the water, lowers 
vthe vapor pressure; now when the light is very feeble it seems probable that 
there still may be a formation of drops of water which, however, in consequence 
of the very small amount of hydrogen peroxide produced by the feeble light, 
never grow large enough to be visible. Thus we may regard the air exposed to 
the ultra-violet light as full of exceedingly minute drops of water; when the 
expansion takes place, the air rushes violently past the drops and we have a 
condition which in many respects is analogous to the bubbling of gas through 
water. However, when air bubbles through water, there is, as Lord Kelvin has 
shown, negative electricity in the air and positive in the water. Thus, when 
the air rushes past the water drops we should expect the air to contain negative 
ions, the positive ions being on the drops; * * * the formation of a fresh 
water surface is accompanied by the liberation of ions; when the cloud of small 
drops is formed by the weak ultra-violet light there is a creation of a new surface 
of water and the probability of the liberation of ions, the positive ions being 
carried by small water drops would only move very slowly under an electric field 
and so might not be cleared out by any field it is practicable to apply. 

Vincent^ has experimented on the clouds formed by ultra-violet light without 
expansion. He finds that some drops are charged positively, others negatively, 
while others are without charge. He was not able to detect the presence of 
hydrogen peroxide. 

^Phil. Trans. A (1899), 192, 403. 

• Conduction of Electricity through Gases. Cambridge, 2d ed. ( 1906) , 255. 

""Proc. Cambridge Phil. 8oc. (1904), 12, 305. 



EFFECT OF TROPICAL SUNUGHT ON THE ATMOSPHERE. 271 

Wilson* showed that with inteixse ultra-violet light, clouds are produced in 
air, or in oxygen, with a minimum amount of expansion, and he. assumes that 
these clouds are primarily deposited on nuclei cpnnected with the formation of. 
hydrogen peroxide. This does not take place when hydrogen is the gas used, 
and Thomson® has shown that, whereas air bubbled through water is strongly 
charged and contains a large proportion of negative ions, with hydrogen gas the 
ionization is exceedingly small. ; 

These experiments seem to me to indicate that the formation of 
hydrogen peroxide is an essential, intermediate step in these reactions. 
When pure water, or a salt solution is exposed to direct sunlight in 
Manila, hydrogen peroxide is formed very rapidly, strong tests being 
obtained after a few hours. H. D. Gibbs/^ of this Bureau, who examined 
the solutions in a great number of reagent bottles exposed in the labor- 
atory for some months to diffuse light, found in practically every case 
considerable quantities of hydrogen peroxide. It is evident that in the 
Tropics, at least in island regions, with the great quantity of vegetable 
growth, the vigorous transpiration of plants, and the large amounts of 
water continually present at the surface of the earth and in the air, 
all the conditions are present when the sun is shining which are neces- 
sary to charge the area of water surfaces, to form peroxide of hydrogen, 
and in general to increase the proportion of ions in the air according to 
the processes which have been outlined above. Whether there is a true 
ionization of the air by tropical sunlight, apart from such a secondary 
ionization, must be determined by further studies. Considerable evi- 
dence is accumulating to show that the tropical sunlight contains more 
intense ultra-violet light than that in temperate zones. Thus, I have 
shown in another paper ^^ that the decomposition of oxalic a^Jid or of 
oxalic acid catalyzed by uranium salts, is very much more rapid in the 
Philippines than in temperate zones, and Gibbs ^^ has shown that the 
coloration of phenol and of aniline takes place much more rapiSly in 
the Tropics than in more northern zones. 

One of the most striking effects produced by ions is the influence they 
exert on the condensation of clouds. I have often noted, in watching a 
steam jet in the open air in Manila, the remarkable way in which, as the 
sunlight strikes it, it becomes dense and beautifully colored, due to the 
interference and diffraction of the light by the small drops of water, 
while as soon as the sun goes behind a cloud, the jet becomes very 
thin, and the colors, of course, disappear. Other conditions being equal, 
on a cloudy day the mountains near Manila can be seen much more 

^ Loc. cit. 
* Log. cit. 

^** Unpublished research. 
^^ This Journal, Sec. A (1910), 5, 

^^This Journal, Sec. A (1910), 3, 361; Ibid. (1909), 4, 133; Ibid. (1910), 
5, 9. 



272 BACON. 

clearly than on one of sunshine, and I do not believe I err when I state 
that all days * of bright sunshine in the Philippines show a decidedly 
hazy atmosphere, as noted by looking at objects at some distance. I 
believe this fact to be due to the ionization of the air by the sunlight 
and the consequent condensation of very minute drops of water around 
the ions so formed. The mountains are most clearly visible from Ma- 
nila at sunrise and at sunset and on days when clouds protect the lower 
atmosphere from the ionizing radiations of the sun. 

It is an important meteorologic question whether direct sunlight can 
produce a cloud in the atmosphere without expansion. Wilson ^^ was 
not able to obtain a cloud in a closed vessel in sunlight with less than 
the normal expansion, 1.25. He points out that the conditions in the 
open air are more favorable to the production of clouds than those in 
a closed vessel, for in the latter the drops might diffuse to the sides 
before they had time to grow to a visible size, while in the atmosphere 
this way of escape would not be open to them. It is an interesting fact 
that the Philippine Weather Bureau in twenty years of daily observation 
has never observed a cloudless day in Manila. In traveling on the sea, 
where disturbing factors due to land surface, vegetation, etc., are absent, 
it is always noted that as one enters the Tropics the air becomes more 
hazy, especially on bright sunshiny days, a fact which I have often ob- 
served and which has been confirmed by me many times in conversation 
with sea captains. As to whether this is due to greater ionization of 
the air by meaus of light in the Tropics, or simply to the fact of there 
being more moisture in the air to condense around ionic nuclei, must be 
determined by future work. I incline to the view that both factors 
cause this increased hazitiess. Apart from the experimental work which 
seems to show a greater intensity of ultra-violet light in the Tropics and 
apart from the more direct incidence of the sun^s rays into the atmos- 
phere of these regions, resulting in less refraction and consequently in 
less loss of the shorter waves of light, there are other theoretical con- 
siderations which seem to point to a greater intensity of ultra-violet light 
for this zone. The absorption of light by water in the air is quite 
marked in the red end of the spectrum and comparatively slight in the 
ultra-violet. The amount of moisture in the air of the Tropics during 
a year is greater than in temperate zones ; also the total amount of radia- 
tion received from the sun in the course of a year is greater in the 
Tropics than in more northern or southern latitudes. Hence, as the 
longer rays corresponding to the red end of the spectrum are propor- 
tionately much more strongly absorbed by the atmosphere in the Tropics, 
and as the total radiation reaching the earth is also much greater than 
in temperate zones, both the proportionate and the absolute quantities of 

" hoc. cit. 



EFFECT OF TROPICAL SUNLIGHT ON THE ATMOSPHERE. 273 

ultra-violet light reaching the surface of the earth in the region around 
the equator must be much greater than it is in those nearer the pole. 

Arrhenius, in his recent papers on the effect of light radiation pres- 
sure in moving cosmic dust, presents several ideas which may be of 
considerable importance in explaining the distribution of sunlight on 
the earth^s surface. The cosmic dust shot out from the sun is charged 
and, consequently, as it approaches the earth's surface, it tends to collect 
toward the magnetic poles, leaving a region around the equator com- 
paratively free from dust. As this fine dust strongly absorbs the blue 
and ultra-violet portions of the sun's spectrum, both in the chromosphere 
of the sun and the atmosphere of the earth, it is evident that tropical 
regions, comparatively free from it, will receive more of these rays than 
regions near the poles. As the distribution of dust around the sun and 
the rate with which it is expelled from the sun are intimately connected 
with the phenomena of sun spots, it seems evident that relations will 
be discovered between sun spots and the ionization of the air, and, 
in general, the rate of reactions carried on under the influence of sun- 
light. 

In connection with my experiments indicating ionization of the air 
by tropical sunlight, a very interesting question arises as to its ionization 
on the different parts of the earth's surface. In the last few years 
hundreds of measurements of this ionization have been made, and, in 
general, it has been ascribed to radioactive substances. This assumption 
has very strong support in many experimental facts, especially as the 
rate of decay of the emanation agrees in many cases with that of known 
radioactive matter. 

Eutherf ord ^* states : 

There can be little doubt that the large part of the radioactivity (ionization) 
of the atmosphere is due to the radium emanation which is constantly diffusing 
into the atmosphere from the pores of the earth. Since radioactivity has been 
observed in the atmosphere at all points at which observations have, so far, 
been made, radioactive matter must be distributed in minute quantities throughout 
the soil of the earth. The volatile emanations escape into the atmosphere by 
diffusion, or are carried to the surface in spring water,^ or by the escape of 
underground gases and cause the radioactive phenomena observed in the atmos- 
phere. 

Rutherford " also mentions that the results point to the conclusion that a 
large part, if not all of the ionization at the earth^s surface, is due to radioactive 
matter distributed in the atmosphere. A constant rate of production of 30 ions 
per second per cubic centimeter of air, which has been observed in the open air 
at the surface of the earth in various localities, would be produced by the 
presence in each cubic centimeter of the air of the amount of emanation liberated 
from 2.4X10~5 grams of radium bromide in radioactive equilibrium. However, 
it is not likely that the ionization of the upper part of the atmosphere is due 

^* Radioactivity. Cambridge. (1905), 523. 
^""Loc. cit. 



274 . . .-. . ,3ACpN^. ^, 

to this cause alone. In orcjer to explain the maintenance of the large positive 
charge which generally exists in the upper atmosphere, there must be a strong 
ionization of the upper air, a condition which may possibly be due to ionizing 
radiations emitted by the sun, " 

However, there are several statements in the recent literature of 
radioactivity which seem to point to the idea that there also may be 
other factors than radioactive processes which help to cause the observed 
ionization of the air. Thus Kutherford and Allan/ <^ working in Canada, 
found that in winter, with the ground frozen and deeply covered with 
snow, and the wind blowing from the north over snow-covered fields, 
there was quit« as much, and, in many cases, more ionization of the 
air than in summer in the same place. 

This fact is difficult to reconcile with the idea of radioactive emanation 
escaping into the atmosphere from the pores of the earth, and is much 
easier to explain on the basis of a sunlight ionization, or of a secondary 
ionization of the air due to snow sublimation and an intermediate forma- 
tion of hydrogen peroxide. 

Elster and Geitel,^^ who made extensive studies on the effects of 
meteorologic conditions on the radioactivity of the atmosphere, sum- 
marized their conclusions in the following table: 



NO. 


Weather. 


Rate of 
leak of 
charge. ! 

1 
3.77 1 

3.18 

8.58 

13,67 


Kate of ! 
leak of 1 
charge. : 


1 


Fog, wind southeast 


2.64 


2 

' 3 


Fine rain, mist 

Clear, air very transparent __ 


3. 02 

9.82 


4 


Sky half overcast, air very transparent 

1 


13. 83 



It is a curious coincidence, if nothing more, that days when the sky 
is about half overcast are days such as No-. 4, when the chemical activity 
of the sun is greatest on reactions like the blackening of silver salts, 
the decomposition of oxalic acid under the influence of uranium salts, 
and on the ionization of the air by tropical sunlight, in short, on 
reactions effected by the ultra-violet portion of the sun^s spectrum. 

If the radium emanation in the atmosphere and the penetrating radia- 
tion due to radium in the earth are together the main causes of the 
ionization of the atmosphere near the surface of the earth, then it might 
be expected that the ionization over the land would be greater than that 
over the sea. The experimental evidence at present obtainable indicates 
that radium is present in sea water to a markedly less extent than in 
the sedimentary rocks on land ; and since the radium emanation decays to 
half its value in five days, the wind is unable to transport the emanation 

*«PM. Mag. (1902), (6), 4, 704. 
^'Ann. d, Phys. (1900), 2, 425. 



EFFECT OF TROPICAL SUNUGHT ON THE ATMOSPHERE. 275 

from land to places on mid-ocean before the activity is materially de- 
creased. 

Boltzmann ^® made daily observatigns on a voyage f rpm . Dover to New York, 
August 21 to 31, 1904. He found that the value of the ionization of the air 
over the ocean does not differ materially from that which obtains in Germany, 
Lapland, or Canada. Eve," on a vqyage from Montreal to Liverpool, June 21 
to 30, obtained practically the same ionization of the air for the ocean as for 
that over the land. He thinks the results may be due to air blown out to sea 
from the land, and to a slower recombination of ions over the ocean, due to a 
lesser quantity of dust particles in the air, but he admits that this explanation 
is not adequate to account for the facts. 

From this evidence and from my own work I consider that other 
factors, which I have outlined above, must be taken into consideration 
in accounting for the ionization of the atmosphere. There is no question, 
from the immense amount of experimental evidence gathered in the last 
few years, but that radioactive substances do play a considerable role in 
ionizing the air, but I do not consider that such radioactive processes 
account for the whole phenomenon of the ionization of the atmosphere. 
Quantites of experimental evidence are continually being brought forward 
to prove that ions are given off in many chemical and physical processes 
other than those known distinctively as radioactive, and such facts would 
naturally be expected from the electronic constitution of matter. 

In connection with the work on the ionization of the atmosphere, I have 
tested a number of substances from the Philippines to obtain an idea of 
the distribution of radioactive substances in this part of the world. In 
working with solids, one metal can was used in which the normal loss 
of potential was 4 to 5 volts per hour. With liquids, another can was 
employed in which the normal loss was 7 to 9 volts per hour. When 
testing various substances, losses greater than these amounts were con- 
sidered to indicate radioactivity. The results are given below. 

1. One liter of water obtained from the boiling lake of the crater of Taal 
Volcano. The water has stood in the laboratory in a tightly stoppered bottle for 
about two years. 



Time. 

(p. m.) 
1.16 
1.26 


Volts. 
169.91 
146.6 


Volts 
per hour. 

146 


1.36 


140.1 


32.4 


2.21 


131.8 


10.4 



The water is active. 

2. One liter of Manila hydrant water drawn from a tap in the laboratory 
February 3 and immediately tested. 



Time. 

(p. m.) 

3.16 


Volts. 
169.3) 


Volts 
per hour. 


3.26 


147.9J 


68 



"P%«. Ztschr. (1906), 6, 132. 
«P/ta. Mag, (1907) (6), 13, 248. 



276 BACON. 

3. Manila hydrant water, February 4. 

Time, 
(p. m.) 

9.03 

9.08 

9.13 

9.23 

9.33 
11.33 

The water is active. 

4. Barium sulphate precipitate formed in the water (Experiment 1) from the 
boiling crater lake of Taal Volcano. 

Time. Volts 

(p.m.) Volts. per hour. 





Volts 


Volts. 


per hour. 


147.9 


115.2 


138.3 


28.8 


135.9 ' 


13.8 


133.6 




132.7 


5.4 


114.9 


6.9 



8.42 144.5 

9.42 134 



.5) 

:.7{ 



9.8 



Slight activity. Agrees well with Experiment 1. 

5. Barium sulphate precipitate formed in 0t5 liter of water from the yellow 
lake of the crater of Taal Volcano. Water which had stood for about two years 
in a well-stoppered bottle in the laboratory. 



Time. Volts 

(p.m.) Volts. per hour. 

1.10 149.91 

2.10 144.5 



5.40 



Very slightly active. 

6. Barium sulphate precipitate from 0.5 liter of water from the green lake of 

the crater of Taal Volcano, collected at the same time as waters Nos. 1 and 5. 

Time. Volts 

(p.m.) Volts. per hour. 

2.101 

2.10 184.91 ^^^ 



3.10 



180.9 



No activity. 

7. Distilled water employed in the usual manner to test apparatus. 

Time. Volts 

(p.m.) Volts. per hour. 

2.47 131.8[ ^^^ 



.81 
.81 



3.02 131 

(a. m.) 
8.22 155.41 2 

8.37 153.6) 

No activity. 

8. One hundred cubic centimeters of rain water collected from the roof of 
the Bureau of Science building after several hours of rain February 11. Evapo- 
rated to dryness in a platinum dish, and the dish tested in the apparatus. 
Measurements were made about one hour after the rain had fallen. 

Time. Volts 

(a.m.) Volts. per hour. 

9.21 131.8. 

9.41 127.6} 



EFFECT OF TROPICAL SUNLIGHT ON THE ATMOSPHERE. 277 



The water is radioactive. 

9. Finely ground Manila copal resin. Such resin is continually being oxidized 
rather rapidly by the air and gives a strong test for hydrogen peroxide. 





Time, 
(a. m.) 


Volts. 


Volts 
per hour. 




9.45 
10.45 


151.51 
146.5 


5.0 


No activity. 








10. Philippine 


clay. 








Time, 
(a. m.) 


Volts. 


Volts 
per hour. 




11.37 


167.11 






(p.m.) 




7.5 




1.27 


152.61 




Slight activity. 








11. Philippine 


clay. 








Time, 
(p.m.) 


Volts. 


Volts 
per hour. 




2.08 


200.81 
194.1! 






3.08 


6.7 


Very slight activity. 






12. Philippine 


clay. 








Time, 
(a. m.) 


Volts. 


Volts 
per hour. 




9.05 


173.31 






10.05 


167.1] 


6.2 



Very slight activity. 

13. Barium sulphate precipitate formed in some rain water collected on 
February 11. 



Time. 

(a.m.) Volts. 


Volts 
per hour. 


11.55 182.0) 
12.55 175.5J 




6.5 


Slightly active. 




14. Powdered gold ore from Benguet. 




Time. 

(p.m.) Volts. 


Volts 
per hour. 


1.05 173.9 
4.05 138.3 


11.9 


Slightly active. 




15. Black sand from Nueva Ecija. This black sand contains magnetite, zircon, 


gold, platinum, and probably iridium and monazite 


!. The decided activity was 


ascribed to the monazite. 




Time. 

(p.m.) Volts. 


Volts 
per hour. 


2.05 173.9) 
3.35 117.4 


37' 


16. Benguet gold ore measured in large pieces. 




Time. 

(a. m.) Volts. 


Volts 
per hour. 


11.15 134.1) 




1.00 118.2/ 


9 



278 



BACON. 



Slight activity. 

17. Black sand from near Paracale, Ambos Camarines. 



Time, 
(a. m.) 

8.40 

10.25 



Volts. 
172.21 
114.9] 



Volts 
per hour. 

33 



Pronounced activity. 

The nonmagnetic portion of this same black, sand. 

Time, 
(a. m.) 

10.20 

10.50 

Very pronounced activity, and grains of monazite were found in this portion. 
18. Black sand from Aroroy, Masbate. Carries a little gold, no platinum. 



Volts. 


per hour. 


201.21 
174.4; 




53.6 



Time, 
(a. m.) 

10.55 

12.55 



Volts. 
153.1 
146 



i\ 



Volts 
per hour. 

10.2 



Slight activity. 

19. Black sand from Ragay Gulf, Tayabas Province, pueblo of Kilbay. 



Time, 
(a. m.) 

11.10 
10.40 



Volts. 
215.0 
209.9 



Slight activity. 

20. Black sand from Rapu-rapu Island, Albay. 



Time. 

(p. m.) 
1.00 
2.00 



Volts. 
148.4 
142.5 



Volts 
per hour. 

10.2 



Volts 
per hour. 



5.9 





Volts 


Volts. 


per hour. 


201.21 




197.4 


5.7 



Very slight activity. 

21. Magnetic portion of black sand from San Mateo, Rizal Province. 

Time, 
(p.m.) 

2.20 

3.00 

Very slight activity. 

Nonmagnetic portion of the same. 

Time. 

(p. m.) Volts. 

3.00 141.3) 

4.00 133.6J 

Slight activity. 

22. Sand from Gumaus River, Ambos Camarines. 



Volts 
per hour. 

7.7 



Time, 
(a. m.) 

9.10 

9.40 



Volts. 
182.9 
180 



::i 



Volts 
per hour. 

5.8 



EFFECT OF TROPICAL SUNUGHT ON THE ATMOSPHERE. 279 

Very slight activity. 

23. Black sand from Ambos Camarines from claims of IngersoW. 

Time, 
(a. m.) 

9.45 * 

10.15 

Slight activity. 

24. Black sand from Lamao Beach, Manila Bay, Bataan^ Province. 





Volts 


Volts. 


per hour. 


167.1) 




167.1 


9.0 



Time, 
(p. m.) 
12.50 


Volts. 
162.61 
152.(1 


Volts 
per hour. 


3.20 


5.3 


Very slight activity. 

Nonmagnetic portion of the 


same. 




Time, 
(p.m.) 

12.00 


Volts. 
195.7) 


Volts 
per hour. 


12.45 
Slight activity. 

25. Black sand collected in 


189.7) ^-^ 
Paracale, Ambos Camarines. 


Time, 
(a. m.) 

9.40 


Volts. 
179.0) 


Volts 
per hour. 


10.10 


176.1 


5.8 



Very slight activity. 

Nonmagnetic portion of the same. 





Time. . 
(a. m.) 

8.30 


Volts. 
139.5) 
131.3) 


Volts 
per hour. 




9.30 


8.9 


Slight activity. 

26. Black sand from Calabasa River, 


Nueva Ecija. 




Time, 
(a. m.) 

8.00 


Volts. 

212.6) 

207.5; 


Volts 
per hour. 




8.40 


6.8 


Slight activity. 

27. Black sand from O'Connell River, 


Tarlac Province. 




Time, 
(a.m.) 

9.45 


Volts. 
216.0) 
192.8 


Volts 
per hour. 




11.00 


7.1 


Slightly active. 








28. Beach sand from Lamao, Manila 
is not black. 


Bay, Bataan Province, at places where it 




Time, 
(a.m.) 

8.45 

10.05 


Volts. 
190.8) 
182.9 


Volts 
per hour. 

5.3 



280 BACON. 

Very slightly active. 

29. Black sand from Ambos Camarines. Exact locality unknown. 

Time. Volts 

(a.m.) Volts. per hour. 

11.16 158.7) 



12.05 154.2 f ^'^ 

Very slightly active. 

30. Black sand from Agusan Eiver (Bukidnon near Cagayan) Mindanao. 

Time. Volts 

(a.m.) Volts. per hour. 

8.35 218.9) 

10.20 207.9] ^ 

Slight activity. 

31. Black sand from bed of Manigui Creek, Agusan Province (Bukidnon), 
Mindanao. 



Time. Volts 

(a.m.) Volts. per hour. 

10.25 208.3) 

11.45 198.3] '^•'^ 



Slightly active. 

32. Black sand from the stream at the base of Mount Panacion, pueblo of 
Linabo, Agusan Province. 



Time. Volts 

(a.m.) Volts. per hour. 

11.45 198.3) 

1.00 188.4] ''^ 



Slightly active. 

These results demonstrate that there is a slight, but rather generally distri- 
buted radioactivity throughout the Philippine Islands, the waters of Taal Volcano, 
and some of the sands containing monazite alone showing pronounced phenomena. 



A SOLUTION OF OXALIC ACID AND URANIUM SALTS AS A 
CHEMICAL PHOTOMETER.' 



By Raymond F. Bacon. 
{From the Chemical Laboratory, Bureau of Sciencey Manila, P. I.) 



The study of tropical sunlight was begun by H. D. Gibbs two and 
one-half years ago in this laboratory. ^ Although the classical researches 
of Bunsen and Roscoe and others have already thrown much light on the 
effects of sunlight in various parts of the world on certain chemical 
reactions and certain differences in the character of sunlight in the 
Tropics and in temperate zones have been discussed by these authors, 
still many interesting facts have presented themselves during the nine 
years of existence of this tropical laboratory, which convince all .of us 
that the subject is still very obscure. Therefore it has been decided to 
take up the investigation from several standpoints, in* order to determine 
in what ways the sunlight reaching the earth in the equatorial zone 
differs from that in other latitudes; preliminary work having indicate^ 
that^ for many chemical reactions the effect of the sunlight is much 
more pronounced in the Tropics than would seem to be indicated by the 
former investigattons of others. Biologic literature is filled with articles 
detailing various effects, mostly deleterious, of tropical sunlight on animal 
arid vegetable life, and there seems to be a general belief that the Tropics 
are not suitable as a permanent residence for white men. One of the 
principal reasons generally assigned for this condition is the effect of the 
sunlight. It is obvious that any experiments made with living forms 
can not be as exact as when nonliving matter is used, as in the former 

^ The results so far obtained on this subject are given in the following paper, 
although the investigation is not as yet complete. The work is being continued 
by Mr. H. D. Gibbs of the Bureau of Science, and Doctor Bacon will make a 
series of comparative measurements in Washington, and the joint results obtained 
by these two investigators, and by others in diflferent parts of the world, will be 
presented later. (P. C. F.) 

*A r6sum€ of some of the aspects of this work is given in the address. of the 
president of the Far Eastern Association of Tropicftl ilSf edicine : Freer, This 
Journal, Sec. B (1910), 5, 1. 

96918 r-4 281 



282 BACON. 

case there is such a large adaptability^ too many conditions which can 
not exactly be controlled^ and, consequently, no one property can be 
selected which can accurately be measured and known to be influenced 
alone by the sunlight. Thus, while it is hoped ultimately to extend this 
investigation to living forms, and while one of the ultimate objects of the 
researches would be to discover the effects of tropical sunlight on man, 
it is necessary first to make an extended series of measurements of the 
effects of this light on the speed of chemical reactions, so that if possible 
it may be ascertained wherein there is a physical difference in the tropical 
sunlight as compared with that of temperate zones. The solutions which 
have been extensively used in this connection as chemical photometers 
all have some grave defects. 

Bunsen . and Roscoe ^ used the combination of hydrogen and chlorine. This 
reaction is exothermic and is also catalyzed by the products of the reaction. 
While Bunsen and Roscoe largely avoided these disturbing factors by the design 
used in constructing the apparatus, the method is still open to considerable error, 
and, furthermore, it is not a convenient one to operate, quite a necessary factor 
when making extended series of measurements in different places. These authors 
also used the blackening of a specially prepared eilver paper. This reaction 
suffers from the disadvantage that ready prepared sensitized papers vary greatly, 
especially in the Tropics where they deteriorate very rapidly, and if special paper 
is freshly prepared the process is somewhat tedious and it is difficult always to 
secure a silver salt of the same degree of sensitiveness. Eder * proposed the 
reaction of mercuric chloride on ammonium oxalate, mercurous chloride being 
separated in the light. As Eder's solution, becomes very cloudy in the light, it 
is obvious that very soon after being exposed to the sunlight, the reaction would 
proceed quite slowly, as little light can penetrate into the solution. Moreover, 
the reaction has a marked temperature coefficient; with a fixed light intensity 
it proceeds regularly in 5, slower and slower measure and is very markedly 
catalyzed "^ by many substances, especially by carbon dioxide, which is one of 
the products of reaction, and hence is never present in constant amount. There- 
fore, we must discard Eder's solution as being unfit for our purposes. Marchand ® 
made a large number of measurements with a solution of oxalic and ferric 
chloride. As I have shown in a previous paper,^ the action of oxalic acid on 
ferric chloride in the light proceeds very rapidly during the first few minutes of 
exposure, but the rate diminishes very soon because of the separation of the iron 
as insoluble ferrous oxalate. 

Duclaux ® used dilute solutions of oxalic acid to measure the effect of sunlight. 

*Ann. d. Phys. u. Ghent. (Poggendorff) (1859), 108, 193; lUd. (1862), 117, 
629; lUd. (1866), 128, 296; lUd, (1874), 151, 268.. 

* Sitzungsher. Akad. d. Wiss. math.-not. Klasse, Wien (1879), 80, 636. Eder's 
solution has 80 grams of ammonium oxalate and 50 grams mercuric chloride in 
3 liters of water. 

■See also on this point Max Roloff, Ztschr. f. phys. Ghem. (1894), 13, 327. 
•Etude sur la force chimique contenue dans la lumiere des soleil. 1875. 
"^This Journal, Sec, A (1907), 2, 129. 

* Atmospheric Actinometry and the Actinic Constitution of the Atmosphere. 
Smithsofir. Gontrih. Knowl. (1903), 29, article 1034. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 283 

I shall show below that the decomposition of oxalic acid, as he used it, is due 
to an intermediate formation of hydrogen peroxide, and is so markedly autocata- 
lytic that over half the decomposition would often take place during the last 
two hours of the afternoon, when the sun's power is quite feeble. Hence, this 
solution must also be discarded, although I made a few measurements with it 
in Manila to compare them with Duclaux's results in France and elsewhere. 

In a previous paper/ in which I made some brief studies on the action 
of uranium salts on oxalic acid to determine whether radioactivit}' was 
connected with the observed effect, I stated that this solution appeared 
to be very promising as a chemical photometer. Further work has con- 
firmed this conclusion, and while it is not ideal from every standpoint, 
still it seems to be the best for the purpose in hand of any which have 
been suggested. 

Oxalic acid^^ solutions absorb almost all of the rays of the ultra-violet 
spectrum, and the same is true of uranium salts, so that the ultraviolet 
portion of the sun's rays may usually be assumed to be that part which 
is aqtive in the reaction. Kayser,^^ in summing up the work on the 
absorption spectrum of uran}^ salt solutions, states that the relations 
are evidently very complicated, but that there is general absorption in 
the violet and ultra-violet regions with bands in the green, blue, and 
violet. Jones and Strong ^^ state in a very recent article that uranyl 
salts in solution are yellow and their absorption spectra consist of a 
broad band of general absorption in the ultra-violet, which extends more 
and more into the region of longer wave lengths as the amount of 
uranyl salt solution in the beam of light is increased. 

Therefore, it may be considered as being fairly well established that 
the active waves from .the sun measured by this solution are in the ultra- 
violet. I have shown in another paper ^^ that there are good grounds 

® Loc. cit. 

^•^ According .to Hartley and Huntington, Phil Trans. (1879), 170, I, 267-274, 
with a 10 per cent solution of oxalic acid absorption begins at wave length 
320 fi fi. According to R. Mazini, Nuovo Ghim. (1903), (5), 6, 343-370, who 
used a layer 1 centimeter in thickness, absorption begins for the different dilutions 
as follpws: 

f 300.9, f„, 270.6, g, 234.4, I 230.0. 

** Handhuch, III, 426. A good summary of the literature is found in Kayser's 
Handhuch III, 418. 

"Awer. Chem. Journ. (1910), 43, 46. No work seems to have been done on 
the infra-red absorption spectra of oxalic acid or uranium salt solutions. See 
Coblentz, Investigations of Infra-Red Spectra. Carnegie Inst. Washington, 1905, 
1906, 1908. The fact shown later that the uranyl acetate-oxalic acid reaction has 
no temperature coejfficient seems to indicate that there are no active rays involved 
in this part of the spectrum. 

'» The Ionization of the Air by Tropical Sunlight with Some Notes on Radio- 
active Phenomena in the Philippines. This Journal, Sec. A (1910), 5, 



284 BACON. 

for believing the tropical sunlight to contain a large proportion of 
such rays^ therefore it was considerect desirable to use a solution sensitive 
to such short wave lengths in the first series of measurements of the 
chemical effects of tropical sunlight. I do not consider that there is 
any good reason for classifying the sun^s rays into infra-red or lieat 
rays, visible rays, and ultra-violet or chemical rays, as there are just as 
many chemical reactions affected by light corresponding to the visible 
and even infra-red parts of the spectrum as there are by the ultra-violet 
portion. 

Since the time of Seek^mp ^* many papers have been published on 
the subject of the decomposition of oxalic acid, but it is rather remark- 
able that none of the investigators have followed the reaction quanti- 
tatively, in order to study the effects of concentration, various catalyzers, 
etc., upon its rate. For practical purposes, the decomposition of oxalic 
acid under the influence of uranyl salts may be assumed to take place, 
as follows; HgCaO^-^COg+CO+HaO. The amount of formic acid pro- 
duced is small, and the quantity of riranyl salt reduced during the 
initial stages of the reaction is also negligible, so that for comparative 
measurements these factors may be disregarded. 

In my earlier work on this subject I used the gas given off as a 
measure of the amount of decomposition of the oxalic acid.. However, 
it is well known that reactions measured by gas evolution are subject 
to errors, due to the size, shape, and material of the vessel, and the 
possibility of supersaturation of carbon dioxide solutions. Hence, in 
my later work, I always determined the undecomposed -oxalic acid, 
acidified with sulphuric acid, by direct titration with potassium per- 
manganate, after removal of the uranium salts with a slight excess of 
ammonium hydroxide. The niranyl salt-oxalic acid solutions were ex- 
posed to sunlight in 100 cubic centimeter Erlenmeyer flasks and the 
rate of decomposition determined by direct titration. 

In the first series of measurements the quantities of oxalic acid and 
uranyl salt were varied within rather wide limits to discover the effect 

^*Ann. Chem. (Liebig) (1862), 122, 113; Ibid. (1865), 133, 253. Fay, 
Amer. Chem. Journ. (1896), 18, 269, gives a good summary of the older work on 
this subject. See also Baur, Ztschr. f. phys, Chem. (1908), 13, 305; Luther, 
Ztschr. f. anorg. Chem. (1907), 54, 1. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 285 



of the concentration of the two substances on the speed of reaction. 
The results obtained are presented in the following tables: 

Table I. — Results with varying amount of oxalic acid and uranyl acetate. 



[Manila, October 21, 1909. 
hazy. Temperature of 
42°-43°.] 



Flasks exposed from 9.05 a. m. to 10.50 a. m. Sky partly 
air, 29 ''-SO**. Temperature of solutions after exposure, 



9_. 
10- 
11- 

12- 
13- 
14. 
15- 
16_ 
17- 
18- 
19- 
20- 
21. 



Number. 



Weight 
uranyl 
acetate. 


Weight 
oxalic 
acid. 


W^ater. 


Found 
oxalic 
acid. 

Oram. 


Oxalic 
acid 
decom- 
posed. 


Gram. 


Gram. 


cc. 


Per cent. 


0.0 


0.5000 


100 


0.495 


1 


0.05 


0.5000 


100 


0.34 


32 


0.10 


0.5000 


100 


0.25 


50 


0.15 


0.5000 


100 


0.16 


68 


0.20 


0.5000 


100 


0.07 


86 


. 0.25 


0.5000 


100 


0.05 


90 


0.30 


0.6000 


100 


0.04 


92 


0.35 


0.5000 


100 


0.04 


92 


0.40 


0!5000 


100 


0.045 


91 


0.45 


0.5000 


100 


0.05 


90 


0.50 


0.5000 


100 


0.05 


90 
Gram. 


0.1 


0.1 


20 


0.14 


0.086 


0.1 


0.2 


20 


0.026 


0.17 


0.1 


0.3 


20 


0.097 


0.20 


0.1 


0.4 


20 


0.140 


0.21 


0.1 


0.5 


20 


0.26 


0.24 


0.1 


0.6 


20 


O.S9^ 


0.20 


0.1 


0.7 


20 


0.486 


0.21 


0.1 


0.8 


20 


0.599 


0.20 


0.1 


0.9 


20 


0.71 


0.19 


0.1 


. 1.0 


20 


0.8 


0.20 



Flasks Nos. 1 to 11, in which the amount of oxalic acid was constant and 
the per cent of the uranyl salt varied, demonstrated that each increment of 
0.05 gram in uranyl acetate content increases the speed of the reaction by 18 
per cent until the limit is reached in No. 5. In flasks Nos. 12 to 21, where the 
amount of uranyl salt is constant and the oxalic acid is varied, except for 
the minimum concentrations of oxalic acid in Nos. 12 and 13 the speed of the 
reaction is constant within the limits of experimental error. These experiments 
show that within very wide limits, the concentration of uranyl salt is the sole 
factor in determining the speed of the reaction. Other experiments giving the 
same general results are as follows: 



286 



BACON. 



Table II. — Results with varying amounts of uranyl acetate solution. 

[One per cent solution of uranyl acetate and 10 per cent solution of oxalic acid. October 
25. 1909. Manila. 11.10 a. m. to 12 a. m. Sky slightly overcast. All solutions 
made up to 20 cubic centimeters.] 



Number. 




1_. 
2_ 
3. 

4_. 
5_. 
6_. 
7_. 
8- 
9-. 
10. 



The effect of axiids, alkalies and a few other substances on the rate of the 
reaction was next studied. The results are presented in the following table. 

Table IV. — Results of the addition of other substances on the rate of reaction. 

[Ten per cent oxalic acid, 1 per cent uranyl acetate solution. Exposed* October 25, 1909, 
11.10 a. m. to 12 m. Sky slightly overcast. All solutions made up to a volume 
of 20 cubic centimeters.] 



Number. 


Oxalic 
acid 
solu- 
tion. 


Uranyl 
acetate 
solu- 
tion. 


Other additions, cubic centimeters. 


Oxalic 

acid 
decom- 
posed. 

Per cent. 
6 25 


11 


cc. 
5 

5 

5 

5 

5 

5 
5 


cc. 

5 

5 
5 
5 

5 

5 
5 


10, sulphuric acid 


12 


10,^ acetic acid 


6 25 


13 . _ . _ 


5, sulphuric acid 


6.31 
6.25 

6.31 

0.0 

6.25 

6.20 

6.25 
6.25 


14 


10, hydrochloric acid 


15 


10, sodium hydrate 


16 _. 


10, 10 per cent sodium hydrate 


17 


1, 10 per cent sulphuric acid 


18 


5, 10 per cent ammonium oxalate 

5, 10 per cent ammonium oxalate and i 

10, ^ sulphuric acid. j 
Control __ - 


19 




20 


5 


5 







OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 287 



Table IV shows that added acids or alkalies have no effect provided they do not 
reduce the content of the solution in uranyl salt, as in number 16, With 
ammonium oxalate the reaction practically proceeds at the same rate as with 
oxalic acid. 

In the following tables the effects of various substances on the speed of the 
reaction are noted, also that of the size of the vessel and the material of which 
it is made. 

Table V. — Effects of various substances^ and of the size and material of the vessel. 

[November 8, 1909. Manila. Sky partly overcast. Ten per cent oxalic acid solution, 
1 per cent uranyl acetate solution.] 



Number. 


Oxalic 
acid 
solu- 
tion. 


Ijranyl 
acetate 
solu- 
tion. 


Time. 

10 50-11.05 
10.50-11.30 
10.45-11.30 
10.45-11.30 

10.45-11.30 

10.45-11.30 

10.45-11.30 
10*45-11.30 


Other additions. 


Remarks. 


Oxalic 

acid 
decom- | 
posed. 1 


2 


cc. 


cc. 


25 cc. Eder's solution, 
10 cc. Eder's solution. 
10 cc. water 




Gram. 

0.0105 
.0931 
.351 
.860 

.01 

.01 

.309 
.267 


3 








4 


5.0 
5.0 

5.0 

5.0 

5.0 
5.0 


5.0 

5.0 
5.0 


Control 


5 


14cc. water, Ice. 5 per 
cent uranyl nitrate. 

15 cc. dilute, solution 
of fluorescein. 

15cc.ditute, solution 
of eosin. 

10 cc. water 




g 




7 




g 


In quartz beaker. 
Flat dish open to 


9 


10 cc. water 






sun. 



Table V shows that there is practically no difference between the nitrate 
and acetate of uranium as accelerators of the reaction, that neither fluorescein 
nor eosin alone have any effect on the decomposition of oxalic acid and that 
the size and shape of the vessel markedly influence the rate of the reaction. 

Table VI. — Effect of various substances, 

[December 14, 1909, 10 a. m. to 1 p. m. Ten per cent solution of oxalic acid, 1 per cent 
' solution of uranium acetate. All made up to a volume of 20 cubic centimeters.] 



Number. 



Solu- 
tion 
oxalic 
acid. 







Oxalic 


Uran- 
ium 


Remarks. 


acid 
decom- 


acetate. 




posed. 


cc. 




Percent. 


10 


Made faintly alkaline tophenolphthalem 
with NaOH. 





10 


Made faintly alkaline to phenolphthalein 
with KOH. 





. 10 


Made faintly alkaline to phenolphthalein 
with NHg solution. 





10 


Control _ 


60 





Flasks 5, 6, and 7 had a decided yellow, opalescent appearance which indicated 
that the uranium was present in some other form than uranyl acetate. 



288 



BACON. 



Table VII. 



[December 15, 1909, Manila, 1.05 p. m. to 3.05 p. m. Solution made to 20 cubic centi- 
meters, 1 per cent uranyl acetate solution.] 



Number. 



Remarks. 



1 5 cc. 10 per cent potassium oxalate solution , 

2 5 cc. 10 per cent oxalic acid solution 



Uranyl 
acetate 
solution. 



10 
•10 



Oxalic 
acid de- 
com- 
posed. 



Gram. 
0.2174 I 
0. 2232 i 



Table VIII. 

[December 16, 1909, Manila, 9.50 a. m. to 12.50 p. m. Ten per cent oxalic acid solution, 
1 per cent uranyl acetate solution. Volumes all made to 20 cubic centimeters.] 



Number. 



1 Ura- ; 
Oxalic I nium 
acid so- acetate 
lution. I solu- I 



Remarks. 





cc. 


cc. 


1 


5 


10 


2_ _ __ 


i 5 • 


10 


3 _ 


■ 1 . 


10 


1 





Made faintly alkaline to phenolphthal- 
ei'n with sodium hydrate solution. 

Made faintly alkaline to methylorange 
with sodium hydrate solution. 

Control i 



Table XI. 



; Oxalic 

; acid de- 

com- 



Per cent. 




[Manila, December 26, 1909, 9 a. 
cent uranyl acetate solution, 
and December 28, 1909, 9 a. 



m. to 12 m. Ten per cent oxalic acid solution, 1 per 
Also repeated on December 27, 1909, 9 a. m. to 12 m., 
m. to 1 p. m.] 



Number. 



a^'cfd so- U^^"-Vl 
?ut1on.!-c«^t«- 



10 

10 
10 
10 
10 
.10 

10 
10 



Remarks 



I Control in [>0 cc. Erlenmeyer 
j flask. 

I In 100 cc. Erlenmeyer flask _. 
I In 400 cc. Erlenmeyer flask __ 

I In quartz crucible 

i In open flat dish 

! 10 cc. water added in 50 cc. 
j Erlenmeyer flask. 

j Trace fluorescein added 

j Control in 50 cc. Erlenmeyer 
flask. 



Oxalic acid decomposed— 



Dec. 26 


Dec. 27 


Dec. 28 


Gram. 


Gram. 


Gram. 


0.268 


0. 251 


0.427 


0. 291 


0.256 


0.476 


0.314 


0.274 


0.476 


0. 302 


0.256 


0.427 


0.403 


0.331 


0.472 


0.314 


0. 261 


0.465 


0.228 


0.182 
0. 248 




0.266 


0. 420 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 289 



Table X. 

[Manila, January 11, 1910, 10.15 a. m. to 1.30 p. m. Sky hazy, 10 per cent oxalic acid, 
1 per cent uranyl acetate. All solutions made up to 20 cubic centimeters.] 



Number. 



1. 

2. 

3: 
4. 

5. 
6. 
7 



Oxalic 
acid. 



Uranyl 
solu- 
tion. 



10 



Remarks. 



Control - 



Added 2 drops concentrated sulphuric acid 

Added 2 drops aniline 

Added 2 drops phenol 

Colored with malachite green 

Colored with methyl violet 



Amount 
oxalic 
acid 
decom- 
posed. 



Grow. 

0.262 

0.0284 

6.0284 

0.008 

0.008 

0.004 

0.004 



Table XI. 

[Manila, November 20, 1909, 10.40 a. m. to 2 p. m. Sun only part of the time. Ten 
per cent oxalic acid solution. One per" cent uranyl acetate solution. All made up 
to a volume of 20 cubic centimeters.] 



Number. 



Oxalic 
acid so- 
lution. 



1 __ _ ___ 


-- i 
cc. 
5 


2 - - - 


5 


3 


5 1 


4 . _ _ 


5 1 


5 


5 i 


6 _ - 


^ 1 


7 


5 ; 


8 .- - 


5 




' 



Uran- 
ium 

acetate 
solu- 
tion. 



10 
10 
10 
10 
10 
10 
10 
10 



Added. 



Control 

0.05 gram quinine 

0.01 gram cinchonine 

0.01 gram strychnine 

0.01 gram brucine 

0.01 gram atropine __ 

0. 01 gram morphine 

0. 5 gram copper sulphate 



Decom- 
position 
of oxalic 
acid. 



Per cent. 
25 
16 
10 
12 
22 
21 
19 
25 



Tables VI, VII, and VIII, show that a certain proportion of hydrogen ions, as 
measured by phenolphthalein, are necessary for this reaction. This is probably to 
be interpreted in the sense that the reaction takes place as long as uranyl salt is 
present. The hydrogen ions per se probably have no effect, as increasing their 
number very markedly did not accelerate the reaction, as shown in Table IV. 

Table IX, in which the same series of measurements was repeated three times, 
demonstrates that increasing the surface of the. solution increases the rate of 
reaction; and also that diluting the solution (No. 6) increases its rate. The 
two observations, Table V, Nos. 8 and 9, which seem to contradict this statement, 
are probably abnormal because of the fact that the solutions became concentrated 
through evaporation. In the experiments given in Table IX, the volumes were 
kept constant to avoid the disturbing factor of change of concentration. In Table 
IX, No. 7, fluorescein was added in the hope that, as it was known to absorb light 
energy, it would assist the uranyl salt in . decomposing the oxalic acid, although 



290 BACON. 

No. 6, Table V, had already shown that fluorescein did not have this effect. Table 
X shows that 0.1 gram of uranyl acetate accelerates the reaction approximately to 
ten times the rate observed with oxalic acid alone, and also demonstrates that 
phenol, aniline, malachite green and methyl violet, like fluorescein, protect the oxalic 
acid from the rays of the sun, probably by preventing the latter from penetrat- 
ing far into , the solution. The experiments of Table XT, prove that small 
amounts of alkaloids act as marked poisons in this catalysis. The results 
agree with those found in my previous paper,^ but it was thought necessary to 
repeat them, using the titration method, first making the solution alkaline, 
and removing the alkaloid with ether before the titration, as it was considered 
possible that in the earlier experiments the alkaloids might have held back 
some of the gas, the volume of which was used as a measure of the amount 
of decomposition of the oxalic acid. The following experiment serves as further 
confirmation of the poisonous character of alkaloids in this instance. 

Two flasks connected with eudiometers were put in the sunlight as follows : 

1. Five cubic centimeters 10 per cent oxalic acid, 5 cubic centimeters 1 per cent 
uranyl acetate solution, 10 cubic centimeters 10 per cent sulphuric acid. 

2. Same solutions with 0.1 gram cinchonine. 

No. 2 gave ^5 cubic centimeters of gas during the time that flask No. 1 gave 
46 cubic centimeters. 

The effect of temperature on the rate of decomposition of oxalic acid by uranyl 
salts is shown by the following: 

1. November 19, 1909, Manila, (a) Five cubic centimeters oxalic acid, 10 cubic 
centimeters 1 per cent uranyl acetate, at boiling temperature. (6) Same solu- 
tion at 30°. ' From 10.46 a. m. to 11.05 a. m. both solutions lost 0.062 gram oxalic 
acid.- 

2. Two flasks each containing 10 cubic centimeters 10 per cent oxalic acid 
and 10 cubic centimeters 1 per cent uranyl acetate solution. Both exposed to 
the sunlight, side by side. Flask A equipped with a reflux condenser and kept 
boiling and Flask B heated only by the sun. Final temperature 44°. A, 0.127 
gram oxalic acid decomposed; B, 0.130 gram. 

These experiments indicate that this reaction over a wide range has no 
temperature coefficient. 

According to Fay,*® Jones found that a very small amount of uranyl oxalate 
will decompose a very large, if not an unlimited quantity of free oxalic acid, if 
exposed continuously to the direct sunlight. 

The experiments which I made on this point are as follows : 

1. Ten cubic centimeters 1 per cent uranyl acetate solution, 10 of 10 per cent 
oxalic acid solution. After a week's exposure to the sun, the green color of the 
solution has entirely disappeared and there is only a grayish precipitate. Titra- 
tion shows practically no oxalic acid left in the solution. In this experiment 0.1 
gram of uranyl acetate has decomposed 1 gram of oxalic acid, and there is no 
indication that the limit has been reached. 

2. Twenty-five cubic centimeters 10 per cent oxalic acid solution; 10 of 1 per 
cent uranyl acetate solution. When the limit was apparently reached, 1.18 grams 
oxalic acid had been decomposed. 

^ Loo cit. 
^ Loc, cit. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 291 

3. Fifteen cubic centimeters 10 per cent oxalic acid solution; 10 of 1 per cent 
uranyl acetate solution. 1.21 grams oxalic acid decomposed. 

These experiments do not agree very well with each other, but they 
seem to indicate that there is a limit to the amount of "oxalic acid 
which a definite quantity of uranyl acetate can decompose, but that 0.1 
gram uranyl acetate can in some cases decompose as much as fifteen 
times its weight of oxalic acid. The reaction is therefore catalytic. 

The effect on the absorption of sunlight by the solution of uranium 
salts on the speed of the reaction was shown by the following: 

A double-walled test tube was prepared, the distance between the two walls 
being at all points 4 millimeters. Ten cubic centimeters of 10 per cent oxalic 
acid were placed in the inner tube together with 10 cubic centimeters of 1 per 
cent uranyl acetate, and the same quantities of these solutions in the outer. The 
whole apparatus was so designed that the level of the liquid in the outer compart- 
ment was about 5 millimeters above that in the inner one; it was exposed to the 
sun for several hours, and then the decomposition in the two tubes measured. The 
amount of oxalic acid decomposed in outer tube, 0.234; in the inner, 0.246 gram. 

The experiment was repeated twice with the same general result, 
namely, the decomposition in the inner tube was always slightly greater 
than in the outer. This seems very surprising and certainly shows 
that there is not a very great absorption of the sunlight in a layer 
of the solution only 4 millimeters thick. It was thought possible that 
the rays, in passing through a thin layer of a solution of uranium salt 
might be so changed as to become more active in exciting the reaction; 
hence a solution of 10 per cent oxalic acid was placed in the inner tube 
and one of uranyl nitrate in the outer. After a week^s exposure to the 
sun, there was practically no decomposition of the oxalic acid. It is 
possible that the observed results may be due to a slight lens effect. 
In any event, the experiment will be repeated in other modifications 
before an attempt is made to draw conclusions. It is interesting to 
note, in this connection, that whereas a solution of uranyl acetate will 
undergo auto-oxidation in the sunlight, and finally deposit a purple 
to black precipitate, the corresponding nitrate in solution seems to be 
perfectly stable to sunlight. Similarly, while a limit of the decom- 
position of oxalic acid by uranyl acetate was apparently reached, some 
preliminary experiments with the nitrate and sulphate indicate that 
with these salts the limit is much higher. 

Bach" states that he obtained formaldehyde and hydrogen peroxide, by the 
action of carbon dioxide on water in the presence of uranyl acetate, a result which 
was strongly ctiticized by Eujer." Nevertheless, it was considered advisable to 

"5er. d. deutschen chem. Qes. (1894), 27, 340. 

^Ihid. (1904), 37, 3414. Bach's answer, loc. oit, 3986; 'ibid. (1906), 39, 
1672. 



292 BACON. 

attempt to confirm these results Under the influence of the tropical sunlight. The 
experiment was performed ^undei- varying conditions, in open flasks and in sealed 
tubeSj and with the acetate, sulphate and nitrate of uranium, but the results in 
every case were negative. No formaldehyde and no hydrogen peroxide could be 
d6tected. In the course of this work it was found that glyoxalic acid (phenyl 
hydrazone melting at 121°) is one of the products formed in the decomposition of 
acetic acid and uranyl salts in the sunlight. 

I was not able to substantiate the fact stated in the older literature ^^ 
that large amounts of formic acid are formed in the reaction of uranyl 
salts oh oxalic .acid solutions^ as is shown by the following typical 
experiment : 

Twenty-five cubic centimeters of 10 per cent oxalic acid to which 25 cubic 
centimeters of 5 per cent uranyl sulphate solution had been added, were exposed 
to the sunlight for several days until the reaction was apparently complete. 
The solution was diluted to a volume of 100 cubic centimeters and 75 cubic centi- 

N 
meters distilled. This distillate was titrated with — sodium hydroxide, using 

phenolphthalein as an indicator, and gave less than 0.002 gram of acid, calculated 
as formic. Titration of the distillation residue proved that oxalic acid had in 
all instances been completely decomposed. Therefore, it is evident that when the 
reaction of uranyl salts and oxalic acid goes to completion,, there is practically 
no formic acid left in the solution. In my former paper ^ I found that formic 
acid was not readily decomposed by uranium salts, hence it appears probable that 
very little was formed during the reaction. However, I will take up this phase 
of the question at a later date. 

Duclaux ^^ made very extensive studies on the actinic power of the 

N . 

sun, using a -j solution of oxalic acid exposed in flat, shallow vessels. 

I shall discuss his results at some length, because, while I do not con- 
sider his solution of great value as a chemical photometer, still his work 
is given in such fonn that comparative measurements under the same 
conditions could be carried on in Manila. This author states: 

"The chemical radiations of the. sun reaching the limits of our atmosphere 
become modified while passing through it, according to a law which is peculiar 
to them, and, «o far as can be seen in so new a subject^ their absorption is not 
the same as that of the calorific or luminous parts of the spectrum." After 
speaking of the fact that landscape photographers often encounter days, on which, 
while the sun is apparently as bright as usual, it is difficult to obtain pictures 
with normal exposures, he continues, "There are days on which for some unknown 
reason, the chemical impression is much slower than on others" and "The very 
fine day of September 15, 1888, gave a combustion of only 9 per cent when the 
slightly veiled day of September 28th gave a combustion of 49 per cent," Thus 

t 

^» Loc. *cit. 

^ Loc. cit. , 

^ Loc. cit. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 293 

May 12, which he classifies as a "fine day," gave a combustion of. 46 per. cent, 
while May 13, given as a "very fine day," gave a combustion of bx^y 29 per cent. 
September 6 and 7 each gave one of 7 per cent, while September 8, classed as a 
"day similar in appearance to the two preceding," gave a combustion of 28 
per cent. 

Many similar figures could be given. Every observer of the effect of 
sunlight on chemical reactions from the time of Bunsen and Roscoe, 
has noted that without regard to the general appearance of the lumi- 
nosity of the sun to the eye, or to the appearance of the sky and the 
transparency of, the air, there are days which must be classed as actinic 
and others as nonactinic. Duclaux considers that so-called "invisible 
actinic clouds of vegetable matter'^ are formed on certain days which 
absorb the actinic rays of the sun and are oxidized, preventing their 
reaching the surface of the earth. This author measured the decomposi- 
tion of oxalic acid in France, in Helsingfors and in Algeria, and found 
that in traveling north, the rate of the reaction becomes greater. He 
says "It (the actinic quality of the sun) would no doubt be found still 
weaker if we approached more nearly to the equator.^^ 

My measurements, made according to Duclaux's methods do not bear 
out this statement, as the rate of decomposition of oxalic acid in. Manila 
is much greater than that which he found for any locality in which 
measurements were made. I consider one of the main factors in causing 
actinic and nonactinic days to be quite different from that assigned by 
Duclaux, for I regard a day with a homogeneous atmosphere as actinic, 
and one during which the atmosphere is more or less stratified as non- 
actinic. It is well known that a comparatively short thickness of air 
will cut out a large amount of the more refrangible rays of light, if 
this air is composed of a series of layers of different densities. The 
measurements of Duclaux in Algeria were made on the edge of a desert, 
where the conditions are ideal for forming columns of heated air and 
ultimately many strata of very different densities. The desert is the 
place par excellence of optical illusions, these having their physical 
basis in such air layers. One would not expect the refrangible, short 
waves of light to reach the earth in quantity in such a region. The 
measurements at Helsingfors were on the seashore, where the great body 
of water tends to make the air homogeneous, while those in the interior 
of France, with a considerable amount of vegetation, would be expected 
to be intermediate between the ones at Helsingfors and Algeria. Trop- 
ical islands almost absolutely covered with vegetation and surrounded 
by the sea offer the least opportunity for the production of columns of 
heated air and therefore stratification, as there are practically no bare 
surfaces of the earth to give rise to heated layers. Moreover, this 



294 BACON. 

tropical region has proved to be very actinic with regard to the decom- 
position of oxalic acid.2^ 

Duclaux states that oxalic acid undergoes a spontaneous decomposition in the 
air, and especially when heated; 10 cubic centimeters of a solution of oxalic acid 
heated for four hours to 95° lost 2.6 per cent and in eight hours 5.2 per cent. 
Neubauer ^ and Bizio ^ both found that oxalic acid decomposes spontaneously. 

N 
I heated 20 cubic centimeters of .^ oxalic acid to between 95° .and 

100° for four hours^.on a stem bath. There was absolutely no loss. 
The older results are probably largely due to the use of poor glassware; 
for Eichardson.^'^ found that oxalic acid is quite stable if kept away 
from light. Solutions of a strength below normal are stable to prolonged 
heating at 100.° 

N 
In the first series of experiments I prepared a ^r. solution of oxalic acid 

which was divided into several portions as follows : 

N 

1. January 15, 1910. One portion of this - oxalic acid was exposed to the 

sunlight for two days in a sealed tube in an atmosphere of hydrogen. In that 
time it had lost nothing and gave no test for hydrogen peroxide. 

N 

2. This portion of the - oxalic acid was exposed in an open Erlenmeyer flask 

during the same time as sealed tube No. 1. The loss of oxalic acid was not 
measurable and the solution gave a strong test for hydrogen peroxide. 

N 

3. This portion of the - oxalic acid was left in a flask which was exposed 

only to the diffuse light of the laboratory. Faint test for hydrogen peroxide. 

.- . . N 

4. This portion was diluted with an equal volume of water forming — oxalic 

acid, which was exposed to the diff*use light of the laboratory as No. 3. 

-^ The influence of sun spots on this phenomena has not yet been, considered. 
It may be a very important factor and in the future work on this subject an 
attempt will be made to correlate known sun spot phenomena with the actinic 
power of the sim as measured by these reactions we are studying. 

At Kieff it has been established that sun spot maxima are coincident with a 
higher average temperature. For the work thus far accomplished in the Tropics 
the results are somewhat contradictory. Koppen ascertained that in the Tropics 
the temperature was by 0.32° lower during sun spot maxima than the average, 
and that five years later, a year before the sun spot minimum, it reached its 
maximum value of 0.41° above the average. However, Very found that the tem- 
perature in very dry districts of the Tropics (near Port Darwin, 12° 28' S., and 
near Alice Springs, 23° 38' S., Australia) is higher at sun spot maxima than at 
minima. Therefore, from Very's measurements, it would appear that the solar 
radiation is really more intense with larger sun spot numbers. Memery also noted 
an instfintaneous rise of temperature immediately a sim spot is first seen ; however, 
probably on account of the ionizing radiation, sun spot maxima are accompanied 
by cloud formation in regions like the wet Tropics, which account for the cooling 
effects noted above. 

^^Ztschr. f. anal Chem. (1870^, 9, 392. 

**/dem. (1867), 6, 62. 

^^Journ. Chem, 8oc. {London) (1849), 65, 457. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 295 



A certain amount of these solutions were now exposed to the sunlight 
in flat Petri dishes on January 15, 1910, from 8.30 a. m. to 4.30 p. m., 
with the following result: 

1. Ten cubic centimeters solution No. 1, 68 per cent decomposition. 

2. Ten cubic centimeters solution No. 2, 96 per cent decomposition. 

3. Ten cubic centimeters solution No. 3, 85 per cent decomposition. 

4. Twenty cubic centimeters solution No. 4, completely decomposed. 

Duclaux says that an old solution of oxalic acid is much more sensitive 
to sunlight than one freshly prepared. He found that fie could sensitize 
his solutions hy exposing them to the sun for a comparatively short time 
and believes this difference to be due to molecular changes taking place 
in the solution which he compares to sensitized collodion, and to the 
variations in the degree of rotation of certain sugar solutions for some 
hours after they are prepared. 

However, my experiments prove that that increase in sensitiveness of 
the solution is due to nothing else than the formation of hydrogen per- 
oxide. When this is not possible, the sunlight has no effect on oxalic 
acid. 

Duclaux also calls attention to the fact that the decomposition of 
oxalic acid at first proceeds very slowly, but continues increasing in rate 
until, during the last two hours of the day, the greatest amount of 
combustion takes place. The following experiments bring out these 
facts more clearly : 

January 12, 13, 1910. Manila. Both rather cloudy days. Two Petri dishes 

N 
each containing 25 cubic centimeters — oxalic acid were exposed from 8 a. m. to 

4 p. m. during both days. One dish with the same solution was exposed January 
12 only and one January 13 only. The quantities of oxalic acid decomposed in 
percentages were as follows: 

January 12 only, 9; January 12 and 13, 69.2; January 13 only, 40; January 12 
and 13, 70.0. 

Therefore, exposed solutions increase in sensitiveness' with the time of 
exposures, just as Duclaux has stated, the reaction b^ing autocatalytic, 
as is shown by the fact that the sum of two exposures on one day is not 
equal to one exposure on two days. 

The following results demonstrate the autocatalytic character of this 
reaction more clearly : 

January 17, 1910, Manila. Bright day., The dishes similar to the above. 

N 
Each dish contained 20 cubic centimeters of — oxalic acid. 



Number. 


Time. 


Decom- 
position. 

Per cent. 
64 
97 
98 


1 


a.m. p.m. 
8.30-2.00 
8.80-4.00 
8.80-4.00 


2 L 


3 





296 



BACON. 



January 18, 1910, Manila. Day somewhat hazy. Each dish contained 20 cubic 
centimeters ^ oxalic acid. 



Number. 


Time. 


Decom- 
position. 


*1 ._ . 


a.m. p.m. 
9.00-1.00' 
9.00-3.30 
9.00-3.30 


Per cent. 
3C 
94 
94 


2 


^3 - 





January 19, 1910, Manila. Day began as a bright one, but there waa not 

N 
much sun during the afternoon. Each dish contained 20 cubic centimeters — 

oxalic acid. 



Number. 


Time. 


Decom- 
position. 


Hours. 


Decom- 
position 
per hour. 


1 _ 


a.m. p.m. 
8.00-10.00 
8.00-11.00 
8.00- 1.00 
8.00- 3.00 
8.00- 4.00 
8.00- 5.00 


Per cent, 

6 
' 12 
24 
48 
76 
100 


2 
3 
5 
7 


Per cent. 
3 


2 


4 


3 , 

4 _ 


5 

7 


5 - - _ ___ 


8 1 9.5 

9 ! 11.1 


6 









The graphic representation given on the opposite page shows the 
results of this autocatalysis very clearly. 

February 4, 1910, Manila. Bright sun for the greater part of the day. Each 

N 
dish contains 20 cubic centimeters — oxalic acid and one dish, containing 20 

cubic centimeters pure water, was also exposed during this time. 



Number. 


Time. 


Decom- 
position. 


Hours. 


Decom- 
position, 
per hour. 


1 . 


a.m. p. m. 
8.40-10.40 
8.40-12.40 
8.40- 1.40 
8.40- 2.40 
8.40- 3.40 


Per cent. 
10 
36 
50 
70 
100 


2 
4 
5 
6 

7 


Per cent. 
5 
9 
10 


2 


3 . „ 


4 


11 


5 _ __ , 


14.3 







A series of duplicate dishes was prepared for examination together 
with this test, and at time of the determination of the acid, tests for 
hydrogen peroxide were also made on these. Number 1 gave a doubtful 
test for hydrogen peroxide, but numbers 2 to 5 show a continually in- 
creasing amount of hydrogen peroxide whereas the dish of pure water 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 297 

gave only a faint test. The results are represented graphically on the 
following diagram. 

rOOx 




3 ^ . 5 6 7 8 9 

Time in hours 

other results with the solutions of oxalic acid in Manila are as 
follows : 

N 



20 



oxalic acid. 



Date. 



Time. 



Decom- 
position. 



a. m. p. m. Percent. 



Remarks. 



February 17 , 8.45-2.30 

Do _■ I 8.45-4.30 

February 23 ' 9.00-2.00 

February 27 i 8.00-3.00 

February 28 | 9.00-4.30 

May 1 L : 10.40-2.40 

March 4 8.30-3.00 

March 8 8.00-4.00 

96918 5 



44 

86 



95 


Bright sun. 


100 


Do. 


91 


! Very cloudy day. 


85 


1 Sun about half the 




time. 


100 


1 Bright sun. 


100 


i Do. 



298 BACON. 

Duclaux, in France, giving an exposure during the hours from 8 a. m. to 
4 p. m., at Mont Dore in August and September, 1888, obtained an average of 
18.3 per cent of decomposition; this special series being lower than usual. A 
decomposition of 50 per cent wa& very usual in France. However, in a series 
from May 9 to June IL, 1888, at Helsingfors, there was, on the average, nearly 
50 per cent decomposition, whereas in Algeria only about 9 per cent was found. 

Other figures in the literature which are of some value for com- 
parison with the Manila results are as follows : 

De Vries^ working in Holland, found from 5 to 25 per cent decomposition 

for 10 cubic centimeters of — oxalic acid in the sun for 8 hours, and in one 

instance 50 per cent. Jorisson and Reicher " studied the decomposition of oxalic 
acid in sunlight, especially with reference to the effect of small additions of 
acids, alkalis and metallic salts. It is impossible to compare their results 
directly with mine for climatologic purposes, as they say they exposed their 
solutions in large flasks, without giving the exact dimensions of the latter, 
but it is evident that the reaction was much slower than in Manila. Thus, in 

N 
one case for seven days of sunlight with 10 cubic centimeters — solution, the 

oxalic acid was decomposed to the extent of only 12 per cent. Jorisson says 
nothing concerning the marked increase in the rate of decomposition with the 
time of exposure, but his results show this phenomenon very clearly. Thus, for 
two days, 15.9 per cent were decomposed, for five days, 100 per cent. 

Days. Per cent. 

1.5 16.7 

2.5 37.5 

4.5 91.7 

1.5 8-7 

3.0 39.1 

5.0 93.5 

Sulc^ working in Prag in the summer of 1898, gives the following figures 

for 10 cubic centimeters — oxalic acid. 

Days. Titre. 

10.04 

. 4 9.92 

19 6.40 

38 0.34 

^ Verslagen en medeelingen der konenklijke Academie von Wetenschaften (188), 
(3), 1, 114. 

''Ztschr. /. phys. Chem, (1899), 31, 142. 
»/5td (1899), 28, 719. 



OXALIC ACI6 AND URANIUM SALTS AS A PHOTOMETER. 299 

N 
In another experiment 100 cubic centimeters — oxalic acid were taken and 

10 cubic centimeters pipetted out each day and titrated. 

Date, 1898. 
June 29 
July 3 
July 9 
July 17 
July 25 
July 31 

Enough data have been given to show that the decomposition of oxalic acid 
proceeds much more rapidly in Manila than in those places in the temperate zone, 
where it has been measured. 

The effect of various other salts on the decomposition of oxalic acid 
was also studied, the results follow. January 19, 1910. Bright sun 
from 9 to 10 a. m. Eather cloudy after 10 a. m. All flasks exposed 
from 9 a. m. to 1 p. m. 



Days. 


Titre. 





11.00 


4 


10.39 


10 


9.40 


18 


7.42 


26 


5.25 


32 


3.62 



Number. 


10 per 
cent 

oxalic 
acid. 


Cubic centimeters of other salts. 


Decom- 
position. 


1 


cc. 
5 
5 
5 
5 
10 
10 
10 
10 
* 10 
10 


10, 1 per cent uranium acetate 


Per cent. 
40 
60 
66 
15 



8 

7 

5 

5 
10 


2 


5, 1 per cent ferrous sulphate Fe804 

5, 1 per cent ferric chloride FeCl s 


3 


4 _ 


10, 1 per cent sulphuric acid 


5 


Trace, copper sulphate 


6 


7 


Trace, manganese sulphate __ _ 


8 


Trace, mercuric chloride 


9 


Trace, chromium chloride 


10 


Trace, cobalt sulphate 







It is evident that many other salts accelerate the decomposition of 
oxalic acid, just as do those of uranium. 

A statement exists in the photochemical literature ^® that platinum, 
gold and iridium salt solutions are reduced more easily by oxalic acid 
which has been exposed to the sun, than by solutions which have not 
been insolated. 

In view of the facts already presented as to the formation of hydrogen 
peroxide by the insolation of oxalic acid solutions, this is easily under- 
stood. 

The older literature contains several statements to the effect that Eder*8 
solutions are subject to a marked period of induction, although a careful reading 



"^Ann. d. Ghent, (liebig) (1862), 122, 193. Draper, Phil. Mag. (1867), 14, 
161, used the separated gold as a measure of the sunlight's action. 



300 



BACON. 



of that author's original paper shows him to liave been aware of the fact that 
such an induction time only represented the period necessary for the liquid to 
become saturated with mercurous chloride. I tested this, using the titration 
method as follows: 

December 15, 1909. Manila. Bright sun during the greater part of the 
experiment. Each flask contained 10 cubic centimeters of Eder's solution. The 
decomposition is represented in terms of cubic centimeters of permanganate, 
containing approximately 0.6 gram per liter. 



Number. 
1 _ _ 


Time(a.m.). 


Decom- 
position. 


lO.OO-lO.lO 
10.00-10.20 
10.10-10.20 
10.00-10.30 
10.20-10.30 
10.00-10.40 
10.30-10.40 
10.00-10.50 
10.40-10.50 
10.00-11.00 
10.50-11.00 
i 


2.0 


2 


4.0 
2.0 
5.0 
2.0 
7.0 
2.0 
9.0 
2.0 
10.6 
2.0 


7 _ _ __. 


3 


8 • 


4 _ _ 




5 


10 -- -- 


6 _ . - 


11 _ . 





There is absolutely no indication of an induction period. In every 
case the speed during the first ten minutes was as great as during any 
of the same periods at a later time. Furthermore, the sum of the 
controls 1, 7, 8, 9, 10, 11 being greater than No. 6, shows that such 
a period does not exist, but that the reaction actually becomes slower 
with time, as Eder has already pointed out. 

A similar series made with a solution of 5 cubic centimeters 1 per 
cent oxalic acid and 5 of 1 per cent uranyl. acetate, demonstrated that 
the reaction with this salt also shows no period of induction and for 
short periods there is no retardation. The sum of the controls was equal, 
within the limits of experimental error, to No. G. Many tests have also 
indicated that there is no hydrogen peroxide formed during the decom- 
position of oxalic acid by sunlight in the presence of uranium salts. 

A few comparative measurements of the action of sunlight on the 
oxalic acid-uranium acetate mixture were made in Manila and in Chicago. 
These indicated, as is shown below, that the sunlight is much more 
active, so far as this one reaction is concerned, here than it is in Chicago. 

Many places in the T^nited States actually have a larger proportion of 
sunlight during the year than the Philippines can show, owing to the 
large proportion of cloudy days in this Archipelago. Thus, the United 
States Weather Bureau gives the percentage of sunshine for the follow- 
ing cities: for JSTew York 56, Pittsburg 44, Atlanta 61, Chicago 
57, Indianapolis 54, Cleveland 45, Cincinnati 38, Santa Fe 76, Salt 
Lake City 62, Los Angeles 7eS, San Francisco 63, Denver 69, as compared 
with 51 per cent for Manila. The larger proportion of sunshine for 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 301 



many cities in the temperate zone tends to a certain extent to equalize 
the lower intensity as compared to Manila. 

Some general comparisons of the action of sunlight in Chicago and 
Manila are brought out in the following table, 100 cubic centimeters 
of solution containing 5 grams oxalic acid and 1 gram uranyl acetate in 
a 100 cubic centimeters Erlenmeyer flask, being used in every case : 

Chicago, May-June, 1910. 



Time. 



12. 20 p. m. to 3. 05 p. m 

12.51 p. m. to 2. 51 p. m 

10 a. m. to 12. 15 p. m 

9. 25 a. m. to 10. 55 a. m 

11 a. m. to 12. 20 p. m 

9.35 a. m. to 10. 35 a. m 

9 a. m. to 11. 30 a. ra 

11.30 a. m. to 1.30 p. m 

1.30 p. m. to3p. m 

3 p. m. to 4 p. m 

Total 



Minutes. 


Cubic 
centime- 
ters of 
gas. 


165 


77 


120 


55 


.135 


85 


90 


51 


80 


58 


60 


15 


150 


43 


120 


62 


90 


58 


60 


86 


1,070 


540 



These figures show that, in the series carried on in Chicago, approxi- 
mately 100 cubic centimeters of gas were evolved in two hundred minutes. 

Manila, October-November, 1909. 



Time. 



10.45 a. m. to 12 n\ 

9. 25 a. m. to 10. 20 a. m 

10. 25 a. m. to 11. 10 a. m 

11.25 a. m. to 12 m 

10.40 a. m. to 11.20 a. m 

1 p. m, to 1.40 p. m 

9. 40 a. m. to 10. 15 a. m 

11.25 a. m. to 12 m 

9. 35 a. m. to 10. 10 a. m 



1 Minutes. 



85 
55 
45 
35 
40 
40 
35 
35 
35 



Cubic 
centime- 
I ters of 
gas. 

i 214 
153 
143 
81 
84 
78 
82 



Total . 



-I 405 



1,018 



Therefore, in the series carried on here, 100 cubic centimeters of gas 
were evolved in forty minutes, or the rate was approximately five times 
that of Chicago. Many more measurements will now be made in tem- 
perate zones and in Manila before we conclude absolutely that the actinic 
power of the sun in carrying on this reaction is so much greater in the 
Tropics than in more northern latitudes. 



302 



BACON. 



Other measurements are as follows : 

U indicates a solution containing 0.5 gram oxalic acid and 0.1 gram uranyl 
acetate in 15 cubic centimeters, E, 10 cubic centimeters of Eder*s solution in 
a 60 cubic centimeter Erlenmeyer flask, and the decomposition is expressed for 
this solution in grams of oxalic acid. 



Date. 



CHICAGO. 

June 3, 1909 

June 4, 1909 

Do 

June 5, 1909 

MANILA 

October 21, 1909 

October 25, 1909 

Do 

October 26, 1909 

October 30, 1909 

November 19, 1909 

Decembers, 1909 

December 16, 1909 



Sol. 


Time. 


Decom- 
position. 


Remarks. 




. 


Per cent. 




U 


9.45-10.30 


1.2 


Bright sun. 


U 


9.00-12.00 


4.5 


Do. 


U 


2.00- 4.00 


1.5 


Sun moderate. 


u 


8.30-12.30 


10.2 


Very good sun. 


u 


9.05-10.50 


48 


Bright sun. 


u 


11.10-12.00 


11 


Sky somewhat overcast. 


u 


2.00- 3.10 


14 


Sun intermittent. 


u 


9.45- 2.30 


20 


Bright sun. 


E 


10.50-11.30 


•0.0931 


Somewhat hazy. 


IT 


10.45-11.05 


12.4 


Moderate sun. 


E 


10.15-10.30 


•0.067 


Bright sun. 


U 


9.50-12.50 


72.0 


Do. 



» Gram. 



Date., 



Sol. 



December 1, 1910 i U 

January 11, 1910 ! U 

January 10, 1910 | U 

Do : E 

Do , U 

Do \ E 

January 11, 1910 ! U 



Do 

January 12, 1910.. 

Do 

Do 

Do 

January 13, 1910_. 

Do 

Do 

Do 

Do .. 

Do 

January 14, 1910.. 
Do- 

January 15, 1910_. 
Do 

January 17, 1910.. 

January 19, 1910.. 



Time. 



9.00-12.00 
9.00-12.00 
8.30- 9.30 
8.30- 9.30 
11.30- 1.00 
11.30- 1.00 
10.15- 1.30 

10.15- 1.30 

9.00-10.00 

9.00-10.00 

11.30- 1.00 

11.30- 1.00 

8.10-11.10 

8.10-11.10 

8.10- 1.10 

8.10- 1.10 

1.00- 3.30 

1.00- 3.30 

10.00- 1.00 

10.00- 1.00 

9.00-12.00 

9.00-12.00 

8.30- 2.00 

9.00- 1.00 



Time. 


Decom- 
position. 


Decom- 
position 
per hour. 


Kemarkg. 


Hrs. 


Per cent. 


Per cent. 




3 


52.6 


17.5 


Sun about half the time. 


3 


50.0 


16.6 


Bright sun. 


1 


6.4 


6.4 


Cloudy. 


1 


•0.0170 


•0.0170 


Do. 


1.5 


31.8 


21.0 


Bright sun. 


1.5 


0.0 


•0.0284 


Do. 


3.25 


52.4 


16.0 


Cloudy about half the 
time. 


3.25 


•0.0170 


•0.0050 


Do. 


1 


25.0 


25.0 


Hazy sun. 


1 


•0.0228 


•0.0228 


Do. 


1.5 


22.8 


15.0 


Sun about half the time. 


1.5 


•0.0228 


•0.0150 


Do. 


3 


60.1 


20.0 


Bright sun. 


3 


•0.0228 


•0.0076 


Do. 


5 


85.0 


17.0 


Do. 


5 


•0.0307 


•0.0076 


Do. 


2.5 


38.6 


15.4 


Do. 


2.5 


•0.0274 


•0.0110 


Do. 


3 


59.2 


19.7 


Do. 


3 


•0.0274 


•0.0091 


Da. 


3 


59.0 


19.6 


Rather hazy day. 


3 


•0.0274 


•0.0091 


Do. 


5.5 


93.0 


16.9 


Bright sun. 


4 


39.2 


9.8 


Cloudy after 10 a. m. 



« Oram. 



OXALIC ACID AND URANIUM SALTS AS A PHOTOMETER. 303 



Date. 



February 4, 1910 

Do 

February 5, 1910 

February 7, 1910 

February 8, 1910 

February 9, 1910 

February 14. 1910 

February 16, 1910 

February 17, 1910 

February 18, 1910 

February 21, 1910 

February 22, 1910 

February 23, 1910 

February 27, 1910 

February 28, 1910 

March 1, 1910 

March 3, 1910 

March 4, 1910 

March 12, 1910 



Sol. 



Time. Time. 



Decom- 
position. 



9.00- 1.00 



Hrs. 
4 



9.00- 4.00 


7 


82.0 


9.00-10.00 


1 


20.0 


8.30- 2.30 


6 


44.0 


9.00- 1.00 


4 


45.0 


9.00- 1.00 


4 


60.0 



10.00- 1.00 ! 3 
10.00- 1.00 I 3 



u 


9.30- 1.30 


i; 


8.45- 1.45 


u 


9.00- 1.00 


u 


9.00- 1.00 


u 


9.00-12.00 


u 


9.00-12.00 


u 


9.00- 4.00 


u 


10.40- 2.40 


u 


11.00- 1.00 


u 


11.00- 1.30 


u 


11.00- 1.00 



7 

4 

2 

2.5 

2 



Per cent. 
64.0 



66.0 
57.0 

90.0 
35.0 
43.0 
66.0 
89.0 
75.0 
91.0 
85.0 
53.0 
66.0 
53.0 



Decom- 
position 
per hour. 



Per cent. 
16.0 

11.7 
20.0 
7.3 
11.2 
15.0 

22.0 
19.0 

22.5 
7.0 
10.7 
16.5 
29.6 
25.0 
13.0 
21.0 
26.5 
26.4 
26.5 



Remarks. 



Sun behind cloud from 
10 to 11. 

Do. 
Bright sun. 
Cloudy dark day. 
Cloudy day. 
Cloudy about half the 

time. 
Bright sun. 

Cloudy about one-quar- 
ter of the time. 
Bright sun. 
Cloudy. 

Cloudy most of the time. 
Bright sun. 

Do. 

Do. 
Cloudy all day. 
Sun about half the time. 
Bright sun. 

Do. 

Do. 



It will be noted that October 26, January 1, January 17, and February 
22, may be classed as nonactinic days, while October 21, November 19, 
December 12, and March 3 and 4 may be classed as very decidedly actinic. 



REVIEWS. 



Theoretical Principles of the Methods of Analytical Chemistry Based Upon 
Chemical Reactions. By M. G. Chesneau, Ing^nieur en Chef des Mines: 
Professeur d'analyse min^rale k TecSle nationale des Mines. Authorized 
Translation by Azariah Thomas Lincoln, Ph. D. Pp. vii+184. Price, $1.75 
net. New York: Macmillan Company, 1910. 

This book is, as its name implies, a general discussion of the processes 
of analytical chemistry viewed in the light of the more modern physico- 
chemical theories. The first two chapters treat of the influence of the 
physical state of precipitates upon their purification by washing, and 
the theoretical principles involved in the processes based upon irreversible 
reactions. The remainder of the book is a study of methods based upon 
reversible reactions by double decomposition of salts, the electrolytic and 
the thermodynamic theories of these reactions being compared by the 
author who decidedly favors the latter. 

While hardly to be recommended as a text-book, this series of lectures 
is well worth reading by teachers or investigators of analytical methods, 
as it gives a rather broad insight into the subject as a whole. 

H. S. W. 



Text-Book of Medical and Pharmaceutical Chemistry. By Elias H. Bartley, 
B. S., M. D., Ph. G. Seventh edition. Pp. vi+734. Price, $3.00 net. Phila- 
'delphia: P. Blakiston's Son & Co., 1909. 

The seventh edition of this book includes some pages on very elementary 
physics, together with .a sketch of systematic crystallography followed by 
General, Inorganic, Organic, and Physiological Chemistry. We believe 
that the time has passed when it is necessary to teach physics to medical 
students in this manner, as all schools fitted to graduate reasonably 
trained physicians should require ample work in this subject before 
admission to the colleges. It is difficult to see how a student can ade- 
quately comprehend modern chemistry without such a knowledge. The 
same statement applies to that portion of the book relating to General 
Chemistry. 

The theories of solution and of ionization and the facts on which they 
are based are treated so sparingly that the student can scarcely be 
expected to be able subsequently to utilize this phase of the science. 

305 



306 REVIEWS. 

The author, in using Berzelius's classification of "metals^^ and "metal- 
loids/^ Berthollet^s "types/^ and other portions of past chemical literature, 
is introducing a conception of the science that, while fundamentally 
useful in its day, has now passed away. 

Some very useful features, such as the physiological effects of com- 
pounds and the study of toxins, are found in the book, and the proteins 
are studied according to the latest classification. Useful data are included 
under the heading of "Food and Diet,^^ which, as a rule, are scattered 
in books on Physiology and Hygiene. The clinical examinations in- 
clude Digestive Secretions, Milk and Urine. The analytical methods 
are briefly and conscientiously explained. 

In my opinion, the medical student had better spend a sufficient amount 
of money to build up for himself a library of special manuals covering 
the subjects dealt with in this book. 

M. V. K. 



JUST ISSUED. 

Order No. 412. 

THE SUdAR INDUSTRY IN THE ISLAND OF NE6R0S 

By HERBERT S. WALKER. 

146 pages, 10 photographic plates, and 1 map. 

Price $1.25, United States currency, postpaid. 



CONTENTS. 



Page. 

Introduction 9 

General information regarding Negros 10 

Geographical location 10 

Size, shape, and area 10 

Mountains 11 

Rivers , 11 

Climate 11 

The sugar belt 14 

The east coast 15 

Other sugar-producing districts of Negros 15 

History of sugar production in Negros 16 

Recent statistics 16 

Varieties of cane grown in Negros 18 

Cane diseases and insect enemies 19 

Nationality of the planters 19 

Native labor: Difficulties, past and present 20 

The principal sugar-producing districts of Negros 22 

Silay 22 

Bago : - 27 

Pontevedra-La Carlota .— - ^. 37 

Binalbagan-Isabela - ,.. 43 

Ilog-Cabancalan 49 

San Carlos ; 55 

Bais 64 

The soil of Negros compared with that of other sugar-producing countries.... 68 

Average composition of the soils of Negros 68 

Hawaiian soils 70 

Egyptian soils 70 

Louisiana soils 71 

Java soils 71 

Demerara soils ^ 71 

Berbice soils 72 

Mauritius soils 72 

Comparison of Negros soils with those of other countries 72 

Fertilization in Negros 73 

[OVBB.] 



Order Ko. 412. 

Page. 

The cane of Negros 76 

Average composition of the purple or native sugar cane in Negros 76 

Other varieties of cane grown in Negros »... 77 

Cane in the Hawaiian Islands^ 79 

Egyptian cane 80 

Java cane 80 

Louisiana cane 80 

West Indian cane 81 

Negros as compared with other countries in respect to the quality of 

cane 81 

Desirability of introducing other varieties of cane 81 

The cultivation of sugar cane and the production of sugar as carried on 

at the present time in Negros 82 

Preparation of the soil 82 

Preparation of the seed 83 

Planting 83 

Cultural operations after planting 84 

Cultivation of ratoon canes 85 

Period of growth of the cane 86 

Cost of cultivation 87 

Cutting the cane 90 

Transporting the cane to the mill 91 

Cost of cutting the cane and transporting it to the mill 92 

Manufacture of sugar from the cane 92 

Extraction of the juice 92 

Manufacture of sugar from the juice 99 

Quality of the sugar produced in Negros 109 

Cost of manufacture 112 

Transportation and sale of the sugar 113 

Estimate of average cost of same 114 

Quantitative experiments to determine the weight of sugar produced 

from a given weight of cane.... 1 v 114 

Mill Control No. 1, hacienda San Jose 117 

Mill Control No. 2, hacienda San Jose 118 

Calculation of the average cost of producing sugar in Negros by the 

methods now employed 123 

Cost for labor alone 123 

Estimate of fixed charges for maintenance and depreciation of plant 

and interest on the capital invested 125 

Total cost of production 126 

Possibilities for improvement 126 

In cultivation 126 

In manufacture 127 

Advantages of a change to modern methods of manufacture 132 

The future of Negros 133 

Summary - 134 

Appendix 139 

An investigation to discover if diseases of the sugar cane exist in 

Negros 139 

Index : 143 

NOTE. 

Orders should be sent to the Business Manager, Philippine Journal of Science, Manila, 
or to any of the below-listed agents. Please give Order Number. 

FOREIGN AGENTS. 

The Macmillan Company, 64-66 Fifth Avenue, New York City, U. S. A. 
Messrs. Wm. Wesley & Son, 28 Essex Street, Strand, London, W. C, England. 
Mr. Martlnus NljhofC, Nobelstraat 18, The Hague, Holland. 

Messrs. Mayer & Muller, Prinz Louis Ferdinandstrasse 2, Berlin, N. W., Germany. 
Messrs. Kelley & Walsh, Limited, 32 Raffles Place, Singapore, Straits Settlements. 
Messrs. A. M. & J. Ferguson, 19 Balllie Street, Colombo, Ceylon. 

CIECUIARS AND DESCRIPTIVE MATTER SENT ON APPLICATION. 



INDO-MALAYAN WOODS 

BY 

Order No. 411, 

FRED W. FOXWORTHY. 

(Being Section C, No. 4, Vol. IV, of the Philippine Journal of Science.) 



182 Pages. 9 Photographic Plates. 

Price $0.50, U. S. currency— Postpaid. 



TABLE OF CONTENTS. 

I. Introduction. 

a. Erroneous popular notions with regard to eastern timbers. 
h. Object of this work. 

c. Definition of the Indo- Malayan region. 

d. Review of previous work and acknowledgments. 

II. Properties of Wood. 

a. Woods of Tropics and Temperate regions compared. 
h. Weight and hardness, tables. 

c. Strength, work of Newton, Gamble and Gardner. 

d. Odor, color, taste. 

III. Suitability of different woods for special purposes. 

a. Enemies of wood. 

h. Woods exposed to salt water. 

c. Ship and boat building, 

IV. Rare, ornamental, or precious woods. 
V. Comparative chart of common names. 

VI. Timbered areas and future supplies of wood. 
VII. Species notes. 
VIII. Index. 



This valuable work is interesting both from the standpoint of com- 
mercial and scientific importance. As will be seen from the Table of 
Contents, the work treats exhaustively of the known Indo-Malayan woods. 

NOTE. 

Orders should be sent to the Business Manager, Philippine Journal of Science, Manila, 
or to any of the below-listed agents. Please give Order Number. 

FGBEION AGENTS. 

The Macmillan Company, 64-66 Fifth Avenue, New York City, U. S. A 

Messrs. Wm. Wesley & Son, 28 Essex Street. Strand, London, W, C. England. 

Mr. Martinus Nijhoff, Nobelstraat 18, The Hague, Holland. 

Messrs. Mayer & Muller, Prinz Louis Perdinandstrasse 2, Berlin, N. W., Qermany. 

Messrs. Kelley & Walsh, Limited, 32 Raffles Place, Singapore, Straits Settlements. 

Messrs. A. M. & J. Ferguson, 19 Balllie Street, Colombo, Ceyloru 

CIRCTTLARS AND DESCRIPTIVE MATTER SENT ON APPLICATION. 

96918 ^6 



Order No. 407. 

THE BATAN DIALECT AS A MEMBER OF THE PHIUFFmE GEOTJP 

OF LANOTJAGES. 

By Otto Soheerer. 

TOGETHER WITH 

"F" AND "V" IN PHILIPPINE LANGUAGES. 

By Carlos Everett Conant. 



141 pages. 



Price $0.80 United States currency, postpaid. 



Order No. 408. 

A VOCABULARY OF THE IGOROT LANGUAGE AS SPOKEN BY THE 

BONTOC IGOROTS. 

By the Eev. Walter Clayton Clapp. 

Igorot-English and English-Igo'rot. 
89 pages. 

Price $0.76, United States currency, postpaid. 



Order No. 102. 



A MANUAL OF PHILIPPINE BIRDS. 

By KiOHARD McGregor. 



Part 1.-412 pages 
Part 11.-324 pages 



> In two volumes. 



Descriptions of all known species of Philippine birds, together with 
keys to species and genera. 

Price (for the two volumes) $4 United States currency, postpaid. 

NOTE. 

Orders should be sent to the Business Manager, Philippine Journal of Science, Manila, 
i>r to any of the below-listed agents. Please give Order Numher. 

EOEEIGN AGENTS. 

The Macmillan CJompany, 64-66 Fifth Avenue, New York City, U. S. A. 
Messrs. Wm. Wesley & Son, 28 Essex Street, Strand, London, W. C, England. 
Mr. Martinus Nijhoff, Nobelstraat 18, The Hague, Holland. 

Messrs. Mayer 6 Muller, Prinz Louis Ferdinandstrasse 2, Berlin, N. W., Germany. 
Messrs. Kelley ft Walsh, Limited, 32 Raffles Place, Singapore, Straits Settlements. 
Messrs. A. M. ft J. Ferguson, 19 Baillie Street, Colombo, Ceylon. 

CmClTLABS AND DESCRIPTIVE MATTER SENT ON APPLICATION. 



THE PHILIPPINE 

Journal of Science 

A. Chemical and Geological Sciences 
AND THE Industries 



Vol. V NOVEMBER, 1910 No. 5 



THE ESSENTIAL FEATURES OF THE GEOLOGY OF THE 
PHILIPPINE ISLANDS. 



By Warren D. Smith. 
{From the Division of Mines, Bureau of Sciencej Manila, P. I.) 



CONTENTS. 
Physiography. 
Topographic Divisions. 

1. Coastal plains. 

2. Intermonta,ne plains. 

3. Intermediate uplands. 

4. Cordilleras. 

5. Volcanoes. 
Stratigraphic and Lithologic. 

1. Tabulated scheme of stratigraphy. 

2. Petrography — andesites, basalts, diorites, etc. 
Geologic History. 

Partial List of Philippine Fossils. 
Economic. 

1. Metals. 

2. Nonmetals, fuels, etc. 

3. Soils. 

4. Artesian water. 

5. Miscellaneous. 

List of the More Important Works Relating to Philippine Geology. 

PHYSIOGRAPHY. 

The Philippine Archipelago forms a link in the great eastern chain of 

islands fringing the Asiatic continent and is a part of the Pacific arc. It 

lies between the J apanese Islands and Formosa on the north, Celebes and 

Borneo on the south, and about 1,000 kilometers east of the mainland. 

99514 307 



308 SMITH. 

The number of large and small islands which it contains is not yet defi- 
nitely known, but- the latest estimate of the Coast and Geodetic Survey 
places it approximately at 3,000. The total area is about 297,917 square 
kilometers, the total length of coast line about 18,532 kilometers. Luzon 
is the largest island, with an area equal to that of England ; Mindanao is 
next in size, with Leyte, Samar, Panay, Negros, Cebu, Palawan, and 
Mindoro composing a secondary group, each of about the area of the 
Island of Crete. There are also a host of lesser islands, some of which are 
nearly as large as those just mentioned, with many others of much smaller 
dimensions down to mere rocks standing out of the sea. 

The Philippines, with Formosa, Japan, Celebes, Borneo, New Guinea, 
and, in fact, the majority of the islands of the Malay group are, in my 
opinion, to be regarded as the rough, shattered ends of the Asiatic conti- 
nent. They mark the border of the continental plateau, the high points 
of the wrinkled continental shelf, the crumpled edge of the great land 
horst. This view seems to be supported by the fact of the existence of 
great ^^deeps'^ found very close to and east of Japan and the Philippines 
during the soundings of the Planet } 

The main tectonic lines in general run north and south, but minor 
departures from this direction are to be seen in the lines running through 
the Islands of Palawan and Mindoro and those in the Sulu group, the 
Zamboanga Peninsula, and Cebu. The various cordilleras, in my opin- 
ion, represent the crests of a series of anticlines the major axes of which 
lie in a north and south line, but the existence of minor cross folding and 
perhaps faulting causes them to be interrupted. The synclinals, the 
bottoms of which fall below sea level, exist as narrow straits common 
throughout the Archipelago. The strait between N"egros and Cebu marks 
such a synclinal. A similar trough, once occupied by water, exists on 
Luzon between the Zambales Eange in the western part of the island 
and the Eastern Cordillera. This is now a long, wide, flat area known as 
the great central plain of Luzon. The lines connecting the various active 
and extinct volcanoes correspond very closely to the main tectonic lines. 

I shall first, in considering the various physiographic units, take up 
the several types of plains, because in them we find the greatest human 
activity. 

COASTAL PLAINS. 

The various units of the Archipelago are, as is the case with many 
recent islands, characteristically lacking in any considerable development 
of coastal plain. The country is backward in commercial development 
probably because of this fact more than any other. Nearly everywhere 
the mountains very closely border the coast. Coral reefs are found 

^Annalen der Hydrographie und Maritimen Meteorologie (1906), Dec, 556. 



GEOLOGY OP THE PHILIPPINE ISLANDS. 309 

fringing the coast line throughout the greater part of its length, and we 
have a coastal plain wherever these have been elevated and covered with 
a veneer of Piedmont deposits. The coastal plains are exceedingly 
broken, thus interrupting the easiest line of communication by land. 
The only stretches of any great importance are from Dagupan north to 
Laoag on the Island of Luzon, the east and west coasts of Mindoro, 
southern Masbate, the Zamboanga plain, some narrow strips on the east 
and west coasts of Cebu, much wider belts on JSTegros, the chief sugar- 
producing island, a portion of the eastern part of Samar, and possibly 
in this connection we should consider the Iloilo plain, although this is 
strictly interment ane. Were there coastal plains of greater extent there' 
would undoubtedly be fewer tribes and dialects. 

THE INTERMONTANE PLAINS. 

The large, more or less flat and fertile tracts which lie between the 
Cordilleras are, with one or two notable exceptions, the sites of the chief 
human activity in these Islands. These plains owe their origin to a 
certain extent to elevation of the troughs between what were once islands, 
and to subsequent upbuilding by delta deposits, alluvial fans, and erup- 
tions of pyroclastic material. The principal plains of this nature are: 

(1) The central plain of Luzon, from Manila Bay to Lingayen Gulf. 
(2) The Cagayan Valley (northern Luzon). (3) The central plain of 
Panay. (4) The valley of the Agusan (eastern Mindanao). (5) The 
Cotabato Valley (central Mindanao). (6) The Albay plain (Luzon). 

The Cagayan Valley, where the greater part of the Philippine tobacco 
is grown, perhaps is the most productive, though not necessarily the most 
fertile plain. The central plain of Luzon is cultivated largely for rice. 
However, much of this land is unproductive and irrigation will go far 
toward reclaiming it. The central plain of Panay is one of the seats of 
the sugar industry. The Cotabato Valley is occupied largely by Maguin- 
danao Moros and produces only a modicum of the crops of which it is 
capable. The valley of the Agusan is as yet practically uncultivated 
excepting for isolated hemp patches. All of these plains were formed 
in much the same manner, and therefore the central plain of Luzon may 
be taken as typical. 

In the beginning there was a structural basin, a deep, wide, and long 
synclinal which was occupied by the sea, giving two islands where now 
there is one, a large one on the east and a smaller on the west, where the 
Zambalas Mountains now rise. Material washing down the mountain 
slopes gradually began to fill up this trough, first in the form of alluvial 
fans and cones and then by means of a broad delta and flood plain which 
slowly progressed southward. In all probability there was some coral 
close to shore, forming a substructure for the plain, but this presumably 
did not* extend very far out into the sea. 



310 SMITH. 

The southern portion of the central plain of Luzon is modified by tuff 
deposits which in places are found in fairly thick beds. The log of a well 
put down near Manila, which is situated at the lower end of this plain, 
shows the character of this material very well. 

Log of well at Pasay, Rizal Province. 

feet to 18 feet. Soil, sand and sea shells. 

18 feet to 36 feet. Gray silt, pebbles, shells. 

36 feet to 50 feet. Gray silt, with shells. 

50 feet to 63 feet. Yellow silt, pebbles, small calcareous concretions. 

63 feet to 83 feet. Yellowish-gray silt, small calcareous pebbles, shells. 

83 feet to 87 feet. Fine to coarse basaltic pebbles and tuff. 

87 feet to 113 feet. Yellow-gray sand, some clay, fragments of soft tuff 
113 feet to 135 feet. Yellow-gray tuff . 
135 feet to 160 feet. Yellow-gray tuff, clayey. . 
160 feet to 180 feet. Yellow sand and tuff, small basaltic pebbles. 
180 feet to 199 feet. Light yellow, gray tuff. 
199 feet to 228 feet. Light gray tuff. 
228 feet to 248 feet. Tuff with fine basaltic pebbles. 
248 feet to 276 feet. Light yellowish gray tuff. 
276 feet to 290 feet. Yellow tuff, small basaltic pebbles. 
290 feet to 318 feet. Tuff and small basaltic pebbles. 
318 feet to 332 feet. Tuff and pebbles from basaltic rock. 
332 feet to 350 feet. Yellowish-gray tuff. 
350 feet to 370 feet. Basaltic pebbles and sand with tuff. 
370 feet to 393 feet. Tuff, small basaltic pebbles. 
393 feet to 422 feet. Tuff. 
422 feet to 432 feet. Tuflf, fine grained. 
432 feet to 463 feet. Tuff, fine grained. 

463 feet to 483 feet. Fine, dark sand, some clear grains, fragments of tuff, basal- 
tic pebbles. 
483 feet to 506 feet. Slightly clayey tuff, basaltic pebbles. 
606 feet to 530 feet. Fine-grained tuff, light gray. 
630 feet to 646 feet. Fine-grained tuff, light gray. 
546 feet to 570 feet. Dark sand, some clear grains. 
670 feet to 594 f«et. Tuff, with small basaltic pebbles. 
694 feet to 620 feet. Clay and basaltic pebbles. 
620 feet to 634 feet. Yellow clay with small basaltic pebbles. 
634 feet to 660 feet. Dark sand, fairly fine. 
660 feet to 690. feet. Dark sand with few basaltic pebbles. 
690 feet to 713 feet. Fine gray tuff. 
713 feet to 743 feet. Basaltic pebbles and fragments of tuff. 

There is an additional feature of this plain not found in the majority 
of the others, namely, there has been a subsequent warping so that two 
streams, the Agno and the Bued, coming out of the mountains from the 
north, are turned westward and finally flow northwest into the sea at 
Lingayen Gulf, while south of this uplift the streams flow south to Manila 
Bay. 



GEOLOGY OP THB PHILIPPINE ISLANDS. lUl 

THE INTERMEDIATE UPLANDS. 

Under this heading I shall consider all the upland territory between 
the foothills and an elevation of 1,500 meters. This includes the rolling 
grass-covered hills of Cebu, Masbate, etc. ; the smaller forest-clad ranges 
of Luzon, the fertile slopes of such volcanoes as Mayon, Canla6n, and 
Ap6, and such semiplateaus as Baguio in Luzon, and the Lanao district 
in Mindanao. 

Only the most primitive people at present live in the uplands, but when 
the mineral and forest resources of the Islands are developed it will be 
found that this zone possibly will furnish the larger portion of the wealth 
of the Islands. The upland zone probably includes 75 per cent of the 
whole area of the Philippines. 

The principal rocks of the upland zone are the Tertiary sedimentaries, 
with some recent volcanics which flank the igneous cones of the Cordil- 
leras. They are found inclined at all angles. However, the sedimenta- 
ries are not encountered everywhere. Some of the upland country, like 
that for instance of a part of northern Masbate, is denuded of sediments 
and characterized by old, worn-down volcanic stocks, while the uplands 
of western Mindanao are covered by a sheet of basalt. All the lode 
mining operations practically are carried on in the upland country. 

Upland limestone country is, contrary to the general opinion, not so 
fertile as that underlain by igneous rock, because the limestone is more 
readily soluble and the rich ingredients are rapidly leached out. Cebu 
Island, which is largely composed of limestone, has a poor soil in the 
upland tracts, but has quite rich patches in the valley areas. On the 
otLer hand, the Lanao upland has an unusually rich soil formed largely 
from the decomposition of basalt. 

Slope also has a great deal to do with the formation of soil. I know 
of several hill districts in the Philippines where the slopes are so steep 
and the rainfall so excessive that soil can not accumulate. The Manca- 
yan-Suyoc district, for instance, is for this reason a country adapted 
neither for cattle nor for agriculture. 

THE CORDILLERAS. 

It is difficult to make a genetic distinction between the intermediate 
uplands and the Cordilleras. It is quite clear that a true cordillera may 
not in its highest parts reach the upper limits I have already placed for 
the intermediate uplands. • For instance, the highest point in the Island 
of Cebu is not much over 1,000 meters above the sea, but the central 
range in this island, which varies from 600 to 1,000 meters in altitude, is 
as wild and uninhabitated as the central range in Luzon, which attains 
to a height of over 2,000 meters, and, technically, it is just as much a 
cordillera. 



312 SMITH. 

- The principal cordilleras are the Sierra Madre Eange (northeast Lu- 
zon), the great Cordillera Central (from Benguet northward to the 
Pacific Ocean), the Zambales Eange (western Luzon), the central ranges 
of, the various Yisayan Islands (Panay, Negros, Cebu, Leyte, Samar, 
Masbate), most of which have only one large central line with two or 
more secondary ones which are lower and therefore come into the category 
of the intermediate upland. There are in Panay an eastern and western 
Cordillera both fairly well developed. The Island of Palawan belongs to 
the Mindoro-Busuanga-Palawan system and is practically cordillera and 
little else. 

A cordillera of only moderate height extends along the Zamboanga 
Peninsula, in Mindanao; a group of fairly high peaks is located in a 
cluster about Lake Lanao; a very irregular group of high points is 
found in southern Mindanao, and a well-defined north and south range 
lies east of the Agusan Eiver; but the continuous and high cordillera 
usually shown on Spanish maps, in line with Mounts Apo and Matutan, 
does not exist. This country for the greater part is quite low, and instead 
of having a well-defined line of mountains it only possesses a few high 
points at wide intervals. The highest altitude on the trail which crosses 
this range is 595 meters. 

As a rule the cordillera in the Philippines is very forbidding country, 
being the home of the most primitive people such as Negritos, Igorots, 
Kalingas, Ifugaos, etc., in Luzon, Manobos in Mindanao, and also the 
haunts of brigands {ladrones as they are called here). (See Plate II 
for the situation of the principal cordilleras.) 

VOLCANOES. 

I have made.no special studies of the volcanoes of the Philippines. 
Mr. H. G. Ferguson, of the division of mines, Bureau of Science, has 
been collecting data for the past three years and intends to publish a 
paper on this subject. Therefore, I have quoted the following from his 
introductory remarks : 

The many volcanoes in the long chain of islands festooning the continent of 
Asia, from Burma to Alaska, for the greater part fall into definite zones, most 
clearly so in the islands of the Dutch East Indies. In the Philippines there are 
also several well-defined belts of volcanoes. Becker ^ has shown that for the 
southern islands of the Archipelago there are two main curved systems intersect- 
ing at an angle of about 60 degrees. He regards one of these lines as containing 
the group of extinct volcanoes forming the Island of Sulu, the Cagayanes Islands, 
and Panay. A second and better marked line, belonging to the same system, 
starts from Darvel Bay in Borneo and contains, in the Sulu Archipelago, many 
recently extinct volcanoes; and in the northward continuation, Tigas and Can- 

«gis* Ann. Rep. U. 8. Oeol. 8urv. (1901), 546. 



GEOLOGY OF THE PHILIPPINE ISLANDS. 313 

laon, the two volcanoes of Negros. The positions of the volcanoes of Mindanao 
relative to the two main fissures are not clear, but it is probable that the line 
of volcanoes extending in northern Celebes through the Sanguir Islands and 
Sarangani to Mount Apo, and perhaps crossing Mindanao to Camiguin de Mi- 
samis, are situated on a parallel outer fissure of this series. Another curve 
parallel to this extends from the eastern side of the Gulf of Davao, through the 
Surigao Peninsula, Leyte, and Masbate, and includes the extinct volcano of 
Diuata in eastern Mindanao and extinct volcanoes in Leyte and Biliran. A 
third outer line, parallel to the last two, is formed by Samar and the Camarines 
peninsula of Luzon. Associated with this are the splendid group of volcanoes in 
Sorsogon, Albay, and Camarines Provinces — ^Bulusan, Bacong, May on, Iriga, and 
Isarog. 

A group of volcanoes is found south of Manila in Laguna and Batangas Prov- 
inces; one of these, Taal, is still active. These do not seem to fit into either 
of the two prevailing systems of the Visayan Islands and southern Luzon, and 
require different schemes for classification. Perry ^ classifies the volcanoes of 
Luzon into three northwesterly lines. The first includes the volcanic stock of 
Mariveles and Taal Volcano; the second, Arayat and Banajao; and the third, the 
Mayon group. Centeno* makes one system to include Arayat, Taal, central 
Mindoro, Canlaon, and Malindang. Koto " omits Arayat and the Taal group from 
his volcanic belt. Becker '' seems to ^include the volcanoes of the Taal region with 
the northwesterly series of fissures and makes no attempt to explain the position 
of Arayat. There are no known volcanoes north of Arayat until the extreme 
northwestern part of Luzon is reached. Northward from Kawa a volcanic chain 
extends due north as far as the Bashi Channel, which separates the Batanes 
Islands from Formosa. 

Three quite distinct types of volcanoes occur in the Archipelago. 
These are : First, the nearly perfect cone ; second, the worn-down stock, 
with no regular form; third, the collapsed cone. Mount Mayon is the 
type of the first group. This is a beautifully symmetrical cone 2,422 
meters in height, the curve of its slope coinciding, according to Becker,^ 

with the hyperbolic sin curve ~ ^=^ ^ ' when c==8.6 mm. The 

C L 

last active eruption of this volcano occurred in 1900. It is an ash cone. 
No lava flows are to be seen anywhere near its base, but recent climbers 
have reported the presence of lava flows near the summit. 

Other volcanoes in this group are Mount Arayat and Bud Dajo, the 
former in Luzon and the latter on the Island of Sulu. 

Mount Mariveles is the type of the second group, and by far the 
greatest number of volcanoes in the Philippines conform to it. 

Taal is the type and the greatest eiample of the third group. 

*Extrait des Annales de la Soci6t6 d'emulation des Vosges (1860), 10, 3d 
pt., 35. 

*Mem. geol.-min. de las Islas Filipinas, Madrid, Tello (1876), 8. 

""Journ. Coll. 8ci. (1899), 11, pt. 2, 112. 

* Loc. cit, 546. 

^ A Feature of Mayon Volcano. Proc, Wash. Acad, 8ci, (1905), 7, 277-282. 



314 SMITH. 

ORIGIN AND AGE OF THE SEVERAL TYPES. 

Mount Mayon is an ash cone. Any violent eruption now would 
probably destroy its perfect outline. Its present slopes bear a definite 
relation to the angle of repose of the materials which compose it. 
Volcanoes of this type are considered to be of very recent age, erosion as 
yet having had but little effect on them. Mayon may be just in its 
prime. 

Mount Mariveles is a type of a volcano which has passed into old age. 
Activity has long since ceased and erosion has made great inroads on it. 
Its great crater has been broken through entirely on one side and a 
great caiion leads out from it toward the sea. This mountain probably 
is a Pleistocene volcano. Both andesite and basalt, chiefly the former, 
and but little ash are found on it. 

Taal Volcano consists of one moderate sized crater, 2,000 meters 
across and 270 meters deep, on a low island in the center of Lake Bombon, 
28 kilometers wide, situated 62 kilometers from Manila. Several smaller 
craters and pools of hot water occur within this large crater. There is 
only one active vent near the center, where a small column of steam and 
gases issues. 

Becker^s idea of the origin of Taal was that the mountain slopes once 
continued upward along a hyperbolic curve, as in Mayon, to an elevation 
perhaps of over 4,000 meters and that a great explosion and subsequent 
sinking of what was left caused the present condition. 

However, Adams'^ theory does not con&ider an explosion at all, nor 
does it presuppose as high a cone as must have existed if we grant 
Becker^s view to be correct. He accounts for the formation of Lake 
Bombon by peripheral faulting with subsidence of the central area. 

There are several minor crater lakes in the Philippines, but that of 
Taal is the largest and best known. Taal was last. in extended eruption 
in 1904. 

STRATIGRAPHY. 

Table 1 gives a tentative scheme of the stratigraphy of the Islands and 
shows that we have no sedimentary formations known with certainty to 
be older than the Eocene; indeed, it is not positively established that 
there are any sediments older than the Oligocene. Nummulites were 
reported by Abella ^ and Richthofen.^^ In their type localities abundant 
orbitoidal forms have been collected, but no nummulites. In the lower 
limestone of Batan Island, Douville found one form which he determined 

'Geological Keconnaissance of Southwestern Luzon. This Journal^ Sec. A 
(1910), 5, 67. 

"Isla de Cebu (1886), 109. 

^""Ztschr. d. geol. Gesell. (1862), 14, 357-360. 



GEOLOGY OF THE PHILIPPINE ISLANDS. 



315 



as Nummulites niasi Verbeek, an Oligocene species in Java. In Ilocos 
Norte and one or two other parts of the Philippines I have found some 
red cherts, or jaspers, which, when examined under the microscope, show 
sponge spicules and fragments of radiolarian tests. These may be equiv- 
alent .to the radiolarian cherts described by K. Martin from the Mo- 
luccas. ^^ 

It is not probable that the igneous complex of diorite, gabbro, etc., is 
very old. At the present time the place in the scale occupied by the 
crystalline schists is entirely a matter of conjecture, for we have only 
lithologic criteria to depend upon. Von Drasche ^^ argued in favor of 
classifying the Agno beds as Paleozoic, but to me the evidence seems to 
be entirely inadequate. There was a time when the greater part if not 
all of the crystalline schists and gneisses were put down as Archean, but 
there is no reason why we can not have metamorphic rocks in the Ter- 
tiary; in fact, such are known. I believe age to be merely incidental, 
the chief factor being dynamism. 

The generalized sections shown in figures 1, 2, 3, and 4, and represent- 
ing northern and central Luzon, Cebti, and Mindanao, respectively, 
graphically illustrate the succession in different parts of the archipelago. 

By far the most predominant rocks are the neovolcanics, andesites, and 
basalts; next come the Tertiary sediments, for the most part Miocene; 
third, the plutonic rocks; fourth, the metamorphics ; and last, the intru- 
sives, such as diabase, quartz-diorites, granites, and dacites. Some 
foraminifera and field notes recently were submitted by me to Professor 
Douville, of Paris, who has studied them and prepared the following 
table :^^ 

Table 2. — The divisions of the Philippine Tertiary {after H. DouvilU) . 



Philippines. 



Borneo. 



i Upper limestone with 
[ small Lepidocyclines. 

Sandstone and shale 



--{ 



Middle limestone with 
large lepidocyclines. 

Lower limestone with 
nummulites, Coal 
Measures. 



Lep. c. f. Verbeeki mio- 
gypsina. 

Cycloclypeus communis, 
Orbitolites alveolinella, 
Miogypsina. 

Lep. insulaenatalis, for- 
mosa, richthofeni. 

Nummulities niasi Verb., 
Amphistegina c. f. niasi, 
Lepidocyclina. 



Burdigalien _ 
Aquitanien__ 

Stampien 



Miocene. 



(Upper Oli- 
l gocene. 



" Reisen in den Molukken, etc. Geologischer Theil. Leiden (1902), 164. 
^Neues Jahrhuch. f. Min. etc. (1879), 265-269. 

" Sur le Tertiaire des Philippines. Compt. rend. Soc. geol. de France (1909), 
14, 130. 



316 



SMITH. 





r 


8 i 


k s 


— *— ' — 1 
S § 




t* 


«o lO 


5^ S 


o 1 


>N 








Q 


? 






-<- 


<0 


•*- 








Qi) 










.? 


<o 










<fi 




























^ 




r * '' 






"? 




V»* 


/. 


« « 






C N 


^ 


•/>}.: 






>) 








«?-v (i. 


« » » « 1 






u 


s • 








C 


\ 








s 


5 

cr 
(0 










c 

< 


^^y* 






Ok 


^ 








1 


■*<^ 


^^^ 






c 


^^S 






.^ 
•^ 



3o 






<o^ 



3 



Vi>iV.JNl 



AVNVl 3 



GEOLOGY OP THE PHILIPPINE ISLANDS. 



317 



PQ 



^'r} 



(D C _ 

C •;: — 



318 



SMITH. 



I- 







GEOLOGY OP THE PHILIPPINE ISLANDS. 



319 



9 § 






^ 






J^ 


» 


•5 


o 


^ 


3 


a 


?; 


i 


a 


03 




fl 


-o 


0) 


o 


^ 


o 








O (U ^ 



11 
" a 






^^ J^ 



o jS 



<1> 08 ^ 



a 






t 






5)H 



a 









320 SMITH. 

PETROGRAPHY. 

The rocks of the Philippine Islands consist of the general groups which 
are common elsewhere. 
These are as follows : 

1. The igneous rocks of the recent volcanoes and the worn-down stocks 
of the older ones, including both extrusives and intrusives. Under the 
intrusives we not only have massive lavas but aerial breccias, or rather 
volcanic agglomerates and tuffs. 

2. The deep-seated igneous rocks, those which have been exposed only 
through long continued erosion. There is every gradation between these 
and the extrusives, and in some cases there may be little or no difference 
in age. These older crystalline massives formerly were thought to be 
very old, perhaps Archean, but there is no good reason for believing 
the majority of them to be of greater age than the earliest Tertiary. 

3. The sedimentary rocks which have been derived by the degradation 
and disintegration of the above-mentioned classes, namely, coral reefs, 
chemicai precipitations, etc. < 

4. The metamorphic rocks such as gneiss, schist, serpentine, etc. 

5. The mineral veins. 

Under the first head we have the following principal types given in 
the order of their predominance : 

(1) Andesites. 

Hornblende andesite. 
Pyroxene andesite. 
Hornblende-pyroxene andesite. 
Olivine-bearing pyroxene andesite. 
Homblende-biotite andesite. 

(2) Basalts. 

(3) Dacites. 

(4) Leucitites. 

The andesites are by far the most common of all of these rocks and 
the pyroxene-bearing variety is dominant. Next to it comes hornblende - 
andesite. The andesites form the older parts of the volcanoes, while 
the basalts generally constitute the later flows. 

A number of petrographers, chief of whom are Oebbeke and Iddings, 
have described various collections of Philippine rocks, and the latter 
has recently worked over the entire collection of the division of mines. 
Bureau of Science. Therefore I shall include some of his descriptions. 
However, as Iddings has made use of several new terms with which only 
petrographers are familiar, I, have taken thei liberty of changing the 
vocabulary in certain instances. These quotations, which are taken from 
a longer and very technical paper already published,^* are printed in 
smaller type. 

^*Thi8 Journal, Sec. A (1910), 5, 155. 



GEOLOGY OP THE PHILIPPINE ISLANDS. 32] 

Pyroxene andesites, — These rocks, as a rule, are dark colored, usually 
dense but often porous, porphyritic, commonly with small phenocrysts. 
Phenocrysts and groundmass occur generally in equal proportions. A 
typical specimen collected at Sisiman, Bataan Province, on the north 
side of the entrance to Manila Bay, is described by Iddings as follows : 

A dark-colored sempatic, mediophyric rock; that is, one having many small 
phenocrysts, about as much in bulk as the groundmass containing them. The 
phenocrysts are mostly labradorite, approximately AbjAns, with pronounced zonal 
structure, the narrow outermost zone being distinctly alkalic. The shapes are 
those of rectangular prismoid to equant crystals. In size they are seriate (that 
is, of different sizes), from those of several millimeters to less than 1 millimeter. 
* * * There are fewer phenocrysts of hypersthene and augite, the former 
faintly pleochroic in thin section. Augite occasionally surrounds hypersthene. 
The pyroxene phenocrysts are euhedral (well-faced ) , with the first and second pina- 
coids strongly developed. They are generally smaller in size than the largest 
feldspar. There is considerable magnetite in small crystals. Those inclosed in 
pyroxene are smaller than others, not 'so inclosed. Some are inclosed in the 
margin of the feldspar. The groundmass consists of microlites crowded together: 
rectangular equant ( equi-dimensional ) , also prismoid plagioclase feldspar, pris- 
moid pyroxene, and equant magnetite; probably with a centimeter matrix of 
colorless glass. 

Hypersthene is very common in many of the Philippine andesites. 

Hornhlende-andesite. — This rock is found in all parts of the Islands, 
forming the summit of Mount Apo and several peaks in the Zambales 
Eange of Luzon. Its habit varies from a rock having large phenocrysts 
of feldspar, 10 millimeters or more in diameter, and smaller ones of 
hornblende, to that in which the relations are just the reverse. They 
are the "trachytes" of the older writers. The disintegration of these 
rocks with the large, glassy plagioclases is the origin of the Orani and 
Tarlac sands, the two best sands for constructional purposes available 
for use in Manila. The layman usually mistakes this plagioclase feldspar 
for quartz^ whereas the sands contain little or no quartz. Professor 
Iddings has described the hornblende andesite used in the Manila break- 
water and which came from Sisiman Point, as follows : 

It is sempatic, seriate, and mediophyric. The most abundant phenocrysts are 
labradorite, AbjAns, euhedral (well-faced) and subhedral (with less perfect faces). 
They possess a narrow outer zone of distinctly more alkalic feldspar, which, 
however, has noticeably higher refraction than the anhedral (without crystal 
faces) feldspar of the surrounding groundmass. The hornblende is greenish- 
brown, but is mostly paramorphosed into aggregates of magnetite and pyroxene. 
There are few phenocrysts of pyroxene tind relatively large ones of magnetite. 
The groundmass is holocrystalline (complete crystal outlines), composed of con- 
sertal (intergrown, anhedrons of feldspar, in part probably orthoclase, with some 
quartz. The rock is somewhat altered in parts, and contains calcite and ( ?) 
chalcedony. 

This is the variety in which the hornblendes are very large and the 



322 SMITH. 

feldspars not so prominent. In the rocks from Mounts Apo and Pinatubo 
the reverse is the case. 

Basalts. — Of this great class of rocks Iddings says : 

There are transitions between olivine-bearing pyroxene andesites and basalts 
rich in olivine, so there are basalts with the texture found in andesite, and others 
with textures not developed in andesites. No line can be drawn between these 
two groups of rocks and petrographers differ as to the classification of rocks 
intermediate between basalts and andesites. It happens that the lavas of Mayon 
and Taal Volcanoes belong in part to these intermediate varieties, which may be 
called olivine-bearing pyroxene-andesites or andesitic basalts, while other varieties 
of lava from these volcanoes are normal basalts, with abundant olivine. 

Basalts with andesitic habit occur in the Batanes. In two cases the rocks 
have the composition of hypersthene-augite-andesite with small phenocrysts of 
colorless olivine partly altered to iddingsite, the groundmass being holocrystalline, 
with anhedral feldspars. These probably are best called olivine-bearing pyroxene- 
andesites. 

Varieties with andesitic habit and much olivine occur at Mount Mariveles in 
Bataan and in Mindanao. In both of these rocks the olivine is present as small 
colorless crystals in part altered to red iddingsite. 

Basalt from the floor of the crater of Taal Volcano is dopatic (groundmass 
dominant), mediophyric (moderately porphyritic), with phenocrysts of subhedral, 
green augite, having inclusions in zones in some crystals, subhedral, equant to 
tabular labradorite, AbjAua — 'Ab, Aua, zonally developed; and fewer colorless to 
yellow olivines, altered on the surface of the crystals. The augite and labradorite 
are anhedral toward each other when in clusters. The olivine in some instances 
is partly inclosed in augite, with anhedral forms. The groundmass consists of 
euhedral prismoid plagioclase, with central euhedral prismoid inclusion that has 
much lower refraction and is isotropic, apparently glass. The plagioclase pris- 
moids have diverse arrangement. There is also much equant anhedral augite, 
less magnetite, and probably intersertal glass, but the microlites are crowded 
close together, and the rock may be holocrystalline. 

Basalt from Mayon Volcano, Albay, is dopatic and mediophyric, and highly ve- 
sicular or porous. The euhedral, rectangular, prismoid phenocrysts of labradorite 
contain many inclusions of brown glass; the phenocrysts of green augite are 
subhedral; those of colorless olivine are subhedral to euhedral. The groundmass 
is dark brown, globolitic glass with microlites of their needle-like prismoids of 
plagioclase, and anhedroifis of augite and magnetite. 

Basalts from Mindanao differ somewhat from those already described from 
Luzon, in being richer in ferromagnesian minerals and in having plagioclase 
slightly less calcic. 

Basalt from the Lanao District, Mindanao, is dopatic mediophyric, and seriate, 
with many phenocrysts of olivine, but slightly altered and with inclusions of 
magnetite. The groundmass consists of much anhedral augite, some anhedral 
olivine, less magnetite, and prismoid plagioclase about equal to the ferromagnesian 
minerals in amount. The composition of the plagioclase is not readily determi- 
nable, it is as calcic as andesine. There is a small amount of colorless matrix 
with lower refraction, which may be glass. 

Leucitites. — ^A dark^ fine-grained rock with small, almost circular white 
areas, 2 or 3 millimeters in diameter, occurs near Aroroy, on the Island 
of Masbate. This proves to be analcite, an alteration product of leucite. 



GEOLOGY OF THE PHILIPPINE ISLANDS. 323 

when examined in thin section. In some cases, though rarely, the original 
leucite can be seen. 

I have seen very similar appearing rocks in northern Luzon, but in none 
of them could I be positive that the small isotropic whitish axeas were 
leucites. The finding of this class of rocks in the Philippines is note- 
worthy, as potash-bearing rocks apparently are of rare occurrence. 

Dacites. — The dacites may be thought of simply as andesites with 
quartz. They are not common, as acidic rocks generally are not common 
in the Philippines. I have seen only two good specimens, one from 
Benguet and one from Corregidor Island at the entrance to Manila Bay. 
For field and petrographic descriptions of this rock the reader is referred 
to Becker ^^ and Iddings/^ respectively. 

The rocks of the second group are in the order of their abundance: 
(1) Diorite, (2) quartz diorite, (3) metadiorites, (4) pyroxenite, (5) 
gabbro, (6) peridotite, (7) granite, (8) syenite. 

1. Diorites. — The parent magma of the Philippines was such that 
when it reached the surface and cooled quickly it produced an andesite 
or dacite; if it did not reach the surface, but cooled slowly at some depth, 
a holocrystalline rock was formed of the same chemical composition, but 
differing in fabric, and diorite was the result. Of course, there were 
variations in the magma so that rocks of just the diorite or andesite 
composition did not always result. However, it was the rule rather 
than the exception. Therefore, we would expect to find the diorites in 
the Cordilleras where the streams have cut deeper into the core of the 
islands. 

The chief characteristics of the diorites are the considerable amount 
of hornblende, and the fact that the feldspar is a plagioclase. When 
they contain quartz, they approach certain varieties of granite; they 
also are closely related to hornblende gabbros. 

The true diorites may sometimes be confused with the metadiorites 
which are abundant, but in which the hornblende is secondary and not 
primary. 

Typic'al diorites have been found in Benguet, Tayabas, Cebu, Ambos 
Camarines, and many other localities. 

2. Quartz diorites have been found in Benguet, Tayabas, and Ambos 
Camarines. Of the Benguet rock (Antamok) Iddings says: 

It is median grained quartz-diorite (grown together) fabric. It consists of 
plagioclase and considerable brownish-green hornblende, anhedral (without good 
faces) with respect to each other but euhedral (with good faces) toward quartz 
and orthoclase. There is some altered biotite. In places the orthoclase is in- 
tersertal (intergrown) to poikalitic with inclusions of plagioclase and hornblende. 

^'^ Geology of the Philippine Islands, 21st Ann. Rep. U. S. Geol. Surv. (1900), 
pt. 3, 516. 
^® Loc. cit. 

99514 2 



324 SMITH. 

other quartz diorites have been encountered in Batangas (Loboo 
Mountains), in Masbate, in Lepanto and elsewhere. 

Metamorphic rocks. — ^These are found in isolated patches in many 
portions of the Archipelago, and are naturally to be seen in or flanking 
the mountainous areas. Some are derived from igneous rocks and some 
from the sedimentaries. Their age is uncertain, many undoubtedly are 
Tertiary. That any of them are Archean is extremely doubtful. There 
is absolutely no reason now for supposing any of those at present 
known or as yet to be discovered to be as old as the Archean. 

We have serp,entines, amphibolites, and magnetite schists in Ilocos 
Norte. In Eomblon, marble and mica schists are found. The marble 
is metamorphosed limestone, and in all probability the mica schists were 
derived from Tertiary sandstones and shales. Some schists of unknown 
origin occur in Cebu on the flanks of the cordillera. In the lower 
Zamboanga Peninsula there is a considerable exposure of quartz-sericite 
schists which may represent metamorphosed sediments, although I am 
not certain of this. In Ambos Camarines we have gneissic granite, 
schistose diorite, slaty shales, and brecciated sandstones. In my opinion, 
we can argue nothing in regard to age from the presence of schists. 
Metamorphics are products of dynamism and may be of any age. 

Owing to lack of space and to the fact that I have described several of 
the Philippine schistose rocks elsewhere,^' I shall not go into greater 
detail at this time. 

3. Metadiorites. — The nietadiorites include a number of rocks in which 
the hornblendes are not primary, but secondary. They are derived from 
holocrystalline rocks containing pyroxene. 

4. Pyroxenite. — Bare, brownish and rocky hills having scattered white 
patches of efflorescence occur in many parts of the Islands. Such hills 
usually are composed of pyroxenite, passing in places to peridotite, and 
where altered it is a serpentine. The efflorescence is magnesite (magne- 
sium carbonate) derived from the decomposition of the ferromagnesian 
minerals in the rock. This rock is almost black, inclined to greenish 
where passing into serpentine, and is very hard and dense. I have noted 
great areas of this formation in Ambos Camarines, Batan Island, Albay, 
and Ilocos Norte. There is usually an accumulation of iron ore associated 
with this formation. 

5. Gablros, — These rocks may be thought of as having cooled too 
slowly to form basalts. They stand in the same relation to basalts as 
do diorites to andesites. This generic relationship of the coarser rocks 
and the extrusive lavas in this region in one of the first things to under- 
stand clearly. The great feature of the quantitative classification is that 

"The Asbestos and Manganese Deposits, etc., of Ilocos Norte. This Journal, 
Sec. A (1907), 2, 145. 



GEOLOGY OP THE PHILIPPINE ISLANDS. 325 

the norms of a basalt and a gabbro are not essentially different and the 
same is true of andesites and diorites. The older classifications did not 
show this. 

As would be expected from the prevalence of basalts in the Archipelago, 
gabbros are quite common. An olivine-gabbro from river gravel at Mod- 
talban, Luzon, contains labradorite, Abg AUg, pale green augite, colorless 
olivine and very small amounts of secondary hornblende, biotite and 
magnetite. There also are some secondary minerals locally developed 
such as chlorite, serpentine, etc, 

A very fine-grained norite is found on Palawan. 

Gabbros with an ophitic texture, that is, with the component minerals 
forming a lattice-like structure, often called diabase or dolerite, are 
numerous. To avoid a multiplicity of names which only confuse and 
lead us away from the true conception of the relationships of these rocks, 
I shall henceforth drop these two terms and keep the name gabbro. 

6. Peridotite. — When a gabbro contains much olivine it is customary 
to call it a peridotite.^® While this rock is not common in the Islands 
there are several occurrences of it. Ambos Camarines is one of the 
localities where it is found. 

7. Granite is a rather rare rock in the Philippines. So far as I know, 
the best development of granite in the Islands is in the Paracale-Mam- 
bulao mining district, Ambos Camarines. It has been so squeezed 
during the regional metamorphism of the district that it now has a 
gneissoid texture. 

Iddings, who visited the district with me, says the granite of Mam- 
bulao has been sheared to a thinly laminated gneiss with "Augen" 
structure on a small scale. The orthoclase and albite lie as anhedral 
blocks in a matrix of smaller equant anhedrous (shapeless) quartz and 
orthoclase with shreds of muscovite (and chlorite, it should be added) 
having pronounced fiuxion structure. 

8. Syenite. — A sodic syenite has been found in Masbate. I have 
seen nothing else which in character even approaches this rock. 

In concluding his paper Iddings says : 

The extremes of this mineral variation, or differentiation in this region, appear 
to be albitic granite and albitic syenite on the one hand, and peridotite and 
possibly pyroxenite on the other. 

There are not sufficient data at hand to determine the relative abundance of 
the diflferent magmas and so indicate the composition of the average or "parent" 
magma. It appears that the coarse-grained, intrusive rocks have attained a 
higher degree of differentiation than the extrusive lavas, but this may not be 
actual fact, and further study of the region may modify this conclusion. 

Sedimentary rocks. — This group includes the usual general classes to 
be found in any part of the world. The following list shows the, order of 

"From the French world peridot tot olivine. 



326 SMITH. 

their abundance: (1) Limestones; (2) shales and clays; (3) sandstones 
and conglomerates; (4) tuff (waterlaid and subaerial) ; (5) cherts. 

Besides these there are numerous subaerial deposits, piedmont depos- 
its, etc. 

1. Limestones, — These are for the most part coralline or foraminiferal. 
They are quite pure chemically, the magnesia content being generally 
less than 2 per cent. They vary in color from cream-white through 
buff to black. They have an extensive development, and vary in age 
from Oligocene to recent. 

According to the foraminifera they contain, Douville has recognized 
three principal horizons, as has been shown by table 2. 

2. Shales and clays. — The former predominate in the. coal measures, 
varying in composition from clay to sandstone and usually are gray, but 
in places may be buff and yellow. They make up many hundred feet 
of thickness of strata in the Philippines. Their greatest development is 
in the Yisayan Islands, where the most extensive coal fields also occur. 

The clays vaiy in composition from very impure varieties high in iron 
content to those which are practically pure kaolin. However, the latter 
are very limited, in fact only small quantities have been found and these 
in La Laguna Province. They are the result of the decomposition of the 
feldspathic volcanic rocks of that region. 

The clays of the coal measures usually contain too much free silica 
and not enough combined silica to be suitable for fire clay or cement 
manufacture. 

3. Sandstones and conglomerates. — In the coal measures there are 
several small seams which I call grit. Sometimes the quartz fragments 
in this grit are over 2 centimeters in diameter. In addition there is at 
least one thick stratum of a very impure, grayish sandstone overlying 
the uppermost coal seams in some parts of the Islands, for instance in 
Cebii. This formation more properly is called an arkose than a sand- 
stone, because it contains more feldspathic, hornblendic, etc., material. 

The conglomerates are of two classes — basal, and those due to consol- 
idation of gravel wash. 

The greatest development of conglomerate in the Islands known to me 
is that bordering the igneous complex of north-central Luzon, as is seen 
in sections along the Bued and Agno Kivers. Other great deposits of 
conglomerate are encountered farther north at an elevation of nearly 
2,000 meters, but they are not basal. 

4. Tujfs. — ^There is a great deposit of pyroclastic material with inter- 
calated beds of silt extending over a considerable area of southwestern 
Luzon, particularly over Cavite and Batangas Provinces and the country 
adjacent to Manila. A great deal of this material probably came fron> 
Taal Volcano, but also from many other vents, many of which in a former 
period existed throughout this region. When first quarried the deposit 



GEOLOGY OP THE PHILIPPINE ISLANDS. 327 

is soft, but it gradually hardens on exposure. It is buff colored to gray. 
It was extensively used in the Spanish regime for building purposes. 
Fragments of pumice, black hornblendes, fragments of feldspar, etc., 
give the rock a very heterogeneous composition. The maximum thick- 
ness of the deposit is probably more than 100 meters. The best exposures 
are to be seen in several quarries along the banks of the Pasig Eiver. 

5. Cherts. — In Ilocos Norte I found a few outcrops of a hard, red, 
fissile chdrt, which on examination proved to have fragments of the tests 
of Radiolaria, These are very similar to some cherts of the Moluccas 
described by K. Martin and assigned to the Jurassic.^® 

As regards distribution, the general statement may be made that the 
oldest rocks probably are those found in the deep canons of the Cordil- 
leras of Luzon. In many other parts of the Islands where we might 
hope to find them we encounter everything covered by a sheet of volcanic 
rocks, as is the case with much of the western part of Mindanao, or else 
by a mantle of coral limestone, as in Cebu. Flanking these older rocks, 
and dipping away from them both to the east and west, are the Tertiary 
sediments, limestone, sandstone, shale, and the intercalated coal seams; 
above these are andesitic and basaltic flows, while the youngest consol- 
idated formation of all is the tuff of the vicinity of Manila. It is not 
easy, in our present state of knowledge, to delimit all of these formations ; 
indeed, many which appear to be of different age are in reality contem- 
poraneous. Another noteworthy fact is particularly well exemplified in 
Cebu, namely, that there is no apparent break in the limestone from 
the coral reef on the shore to the capping of the cordillera in the center 
of the island, at a height of 1,000 meters. I have walked over a limestone 
formation on that island, which is continuous from the living coral reefs 
to the Pliocene, and probably the Miocene, with apparently' no un- 
conformity. This island must have suffered a long period of erosion 
after the Miocene. It probably sank below sea level and subsequently 
rose so gradually that the whole island was covered with a mantle of 
coral limestone. This mantle has since largely been removed by erosion. 

GEOLOGIC HISTORY. 

In taking up this part of my discussion, I feel that the dominant 
episodes in the geologic history are best given in Becker's admirable 
summary and by adding such comments as later and more intimate 
acquaintance with the field shows to be necessary. Mr. Becker says : 

It would seem that the geological history of the Philippines is something as 
follows: From early Paleozoic times onward an archipelago has usually marked 
the position of these islands. Prior to the Eocene nothing definite is known of 
them, but further investigation will very likely disclose Paleozoic and Mesozoic 
strata there, as in the Sunda and the Banda islands. During the Eocene it is 

"Reisen in den Molukken, etc. Geol. Theil. Brill, Leiden (1902), 170. 



328 SMITH. 

probable that the lignitic series of Cebu was deposited, and the contorted in- 
durated strata, which in other localities also carry black lignite relatively free 
from water, should be referred provisionally to this period. Whether the num- 
mulitic limestone found at Binangonan is Epcen6 seems to me to be an unsolved 
question. After the Cebtian lignitic epoch a great uplift and folding took place, 
and this may have been a detail of the late Eocene movement which so profoundly 
modified Asia and Europe. It must have brought about temporary continuity of 
land area between Borneo and Luzon. Somewhere about the middle of the 
Miocene the coimtry^ sank to a low level. Many of the present islands must then 
have been far below water, while Luzon and Mindanao were represented by 
groups of islets. Observations appear to suggest that the Agno beds represent 
the basal conglomerate formed at this subsidence. A slow rise began again dur- 
ing the later Miocene, and may have continued to the present day without in- 
version, yet the actual distribution of living forms is such as to give some 
ground for believiilg that, at some intermediate period, the Islands were a little 
higher than they now are, but sank again only to rise afresh. The diorites and 
associated massive rocks, including their tuflfs, may have made their appearance 
about the close of the Paleozoic. The less siliceous of these rocks seem to have 
followed the more siliceous intrusions as a whole, l" he gold deposits, and perhaps 
other ores, are so associated with these massive rocks as to indicate a genetic 
relation. The neovolcanic period began as early as the highest Miocene horizon, 
and very probably at the Post-Eocene upheaval. If the semiplastic marls of Cebti 
are all Miocene, the earlier andesitic rocks, at least, date back nearly to the great 
upheaval. Among these rocks, also, there is sometimes a tendency for the basalts 
to follow the andesites, but the one dacite found at Corregidor is later than the 
andesites of that island. The relation of the trachytes to the andesites is not 
certain, but the sanidine rock is probably the earlier. A very large part of the 
neovolcanic ejecta has fallen into water and been rearranged as tuffaceous plains. 
The volcanic vents appear to me to occur rather on a network of fissures than 
on a system of parallel diaclases, and the volcanic activity is to be regarded as a 
thermal manifestation of the energy of upheaval. 

As regards the history of* the Islands prior to the Eocene, our knowl- 
edge is in the same^ state to-day as it was when Becker wrote his summary. 
Furthermore, I am not as hopeful as I was at the beginning of my in- 
vestigations that Mesozoic and Paleozoic strata will be found. They may 
by chance be encountered in deep wells, in a deep shaft, or in some canon 
of the Cordillera on the Island of Luzon. The oldest fossil I have seen 
from the Philippines is NummuUtes niasi, which is typically Oligocene. 
Lepidocyclina, Cycloclypeus, Glohigerina, Heterostegina, Lithothamnium, 
etc., are abundant. . The Binangonan limestone is Oligocene. 

It is probable that the Islands have sunk at least once after the 
formation of the Agno beds (Bued River conglomerate). This fact is 
•attested by the presence of an indurated coral reef above the latter at a 
present elevation of ov^r 1,300 meters. This old reef is exposed in 
Trinidad Gap a few miles north of Baguio. It seems quite probable 
that the main period of ore deposition followed the Miocene uplift and 
ended before that of the later sedimentaries. The gold-bearing veins 
in Masbate cut across andesites and diorites indiscriminately. They 



GEOLOGY OF THE PHILIPPINE ISLANDS. 



329 



occur in Benguet, as far as we know, as fissure yeins in the diorite-andesite 
and in the sedimentaries. However, ore deposition may still either be 
going on in Benguet or else have only recently ceased. N"o connection 
between the intrusives and the ores is known with certainty, excepting 
that they both follow (in Benguet) east and west or north and south 
lines of weakness. 

The "trachytes^^ referred to are andesites and therefore are to be 
regarded simply as variations of the andesite-diorite parent magma. 

PARTIAL LIST OF GENERA AND SPECIES OF PHILIPPINE FOSSILS. 



Area nodosa Mart. 
Area hispida PhiL 
^quipecten sp. 
Alectryonia folium Lam. 
Amusiopecten burdigalen- 

sis var. elongata Sacco. 
Astarte sp. 
Balanophyllia sp. 
Bulla sp. 

Callianassa dijkj Mart. 
Campanile sp. 
Capulus sp. 
Cassis pila Reeve. 
Cassis nodulosa Gmelin. 
Cassis herklotzi ( ? ) Mart. 
Cardita boettgeri Mart. 
Cardium rugosum Lam. 
Cerithium karangense 

Mart. 
Cerithium (Vicarya) cal- 

losa Jenk. var. semperi 

Mart. 
Chama sp. 
dementia papyra ( ? ) 

Gray. 
Conus sinensis Low. 
Conus sulcatus Reeve. 
Conus vimineus Reeve. 
Conus odengensis Mart. 
Conus acutangulus ( ? ) 

Chem. 
Conus stritellun Jenk. 
Corbula tunicata Hinds. 
Cultellus m a X i m u s 

Gmelin. 
Cycloseris hidalgi Smith. 
Cyprsea poraria ( ? ) Lim. 
Cyprsea paniculus ( ? ) 

Bttg. 



Cytherea javana Mart. 

Dentalium sp. 

Dictyrsea ( ?) micantha 

Reuss. 
Dione sp. 

Dolium costatum Desh. 
Dosinia boettgeri Mart. 
Dosinia plana (?) Reeve. 
Drillia ( ? ) weberi Smith. 
Euthria sp. 
Fusus cortada sp. nov. 

(?) 
Fusus tjidamarensis 

Mart. 
Fusus (Cyrtulus) ver- 

beeki Mart. 
Glossostylus picturatus 

Pfeifer. 
Gryphsea sp. 
Hindsia dijkj Mart. 
Lagunum ( ? ) . 
Lepidocyclina richthofeni 

Smith. 
Lucina sp. 
Lucina (Codakia) sem- 

periana ( ?) Issel. 
Macoma rosareana Smith. 
Madracis sp. 
Melanea laterita Lea. 
Melanea denticulata Lea. 
Melanea woodwardi Mart. 
Morio echinophora Linn. 
Nassa verbeeki Mart. 
Nassa cassiculata Lam. 
Natica globosa Chem. 
N a t i c a marochiensis 

Gmel. 
Natica rostellina (?) 

Jenk. 
Natica ovum Mart. 



Natica callosior Mart. 
Nerita punctata Smith. 
Nummulites niasi Ver- 

beek. 
Odontocyathus sp. 
Orbitoides (Lepidocyclina 

martini Smith ) . 
Ostrea sp. 

Pachyseris cristata Mart. 
Pecten subarcuatus ( ? ) 

Bttg. 
Pecten reticulatus Reeve. 
Pecten leopardus ( ? ) . 
Pecten senator ius Gmel. 
Pecten Javanus ( ? ) Mart. 
Phos. acuminatus Mart. 
Placuna placenta Linn. 
Pleurotoma gendinganesis 

Mart. 
Pleurotoma carinata 

Gray. var. woodwardi 

Mart. 
Pleurotoma nodif era ( ? ) 

Lam. 
Plicatula imbricata 

Menke. 
Potamides palustris Linn. 
Potamides herklotzi ( ? ) 

Mart. 
Psammobia tenuis Desh. 
Pteropsis (?) bullata 

Smith. 
Ranella spinosa Lam. 
Ranella nobilis Reeve. 
Ranella subgranosa Beck. 
Rostellarea javana Mart. 
Septaria arenarea Lam. 
Semele (?) dalli Smith. 
Stephanocoenia sp. 



330 



SMITH. 



PARTIAL LIST OF GENERA AND SPECIES OF PHILIPPINE FOSSILS — ctd. 



Mart. 



(?) 



Strombus javanus 

var. semperi Sm. 
T a g e 1 u s coribaBUs 

Lam. 
Tagelus coartatus Gmelin. 
Tapes lobooensis Smith. 
Tellina plicata (?) Va- 

lenc. 
Teridina annulata ( ?) 

Bttg. 



Tritonidia ventricosa 

Mart. 
Trochus sp. 
Turbo borneSnsis Bttg. 
Turbo nivosus Reeve. 
Turicula jonkeri Mart. 
Turritella terebra Lam. 
Turritella cingulifera 

Sow. 



Venus pulcherrima ( ? ) 

Mart. 
Venus squamosa Lam. 
Vermetus junghulmi 

Mart. 
Vicarya callosa Jenk. var. 

semperi Sm. 
Voluta pellis serpentis 

(?) Linn. 



Save for some fragments of deer ajid shark teeth found by me in 
the tuff formation in Batangas Province and in the recent tuff, by 
Adams, no fossils of mammals and reptiles, large or small, or of fish 
remains have been found. There are three reasons for this scarcity: 
First, the mammalian and reptilian indigenous fauna was probably 
meager compared with that of the continental tracts, it is even very 
poor to-day; second, there are few excavations in places which would 
most likely 'contain such remains; and, third, no search has, until 
recently, been made for them. 

The coal measures have furnished very few fossil plants, and all that 
I have found can be identified by the botanists at work oi^ the living 
flora. I should be inclined to express the opinion, warning the reader 
that it is but an opinion and based on very fragmentary knowledge, that 
since the middle of the Tertiary, at which time these Islands arose from 
the sea, there has been little climatic change. 

In fact, the tree fern and the vast nipa swamps indicate that these 
Islands are about in the stage of the British Isles in the Miocene period ; 
that is, we are still in the Tertiary period. 

ECONOMIC. 

During the three centuries of Spanish occupation very little mining, 
as we now conceive of it, was carried on. Thousands of Chinese, Fili- 
pinos, and Moros have made wages through desultory panning for gold; 
the semi-wild Igorots of northern Luzon have mined and smelted copper 
and made crude implements ; and one Filipino woman has . operated for 
a number of years crude blast furnaces for smelting iron from which 
equally crude plow shares are made and used by the natives in central 
Luzon. 

Just prior to the insurrection of 1896^ several large mining enterprises 
had just been undertaken with Spanish and English capital and engineers. 
The chief of these was the Philippine Mineral Syndicate. However, 
these were nipped in the bud by the opening of hostilities. It seems 
that practically the only mining works which came to any fruition were 



GEOLOGY OF THE PHILIPPINE ISLANDS. 331 

the numerous arrastres set up and operated in Ambos Camarines by 
Spaniards from Mexico. 

The year 1905 practically marked the beginning of production in the 
Islands under the American regime, that of gold amounting to 35,000 
pesos (17,500 dollars United States currency). The mineral production 
for 1909 is given in Mineral Resources of the Philippine Islands, Bureau 
of Science, Manila, 1910. 

THE METALS. 

Gold. — This metal has been found in some quantity in nearly every part 
of the Archipelago. It was mined in a crude way as long ago as we 
have any records. I have been in ancient workings on the Islands of 
Masbate which probably were made by Chinese years before the coming 
of the Spaniards. Magellan heard reports of gold in the Camarines when 
he reached the Islands in 1521. Although it is not the most important 
mineral asset of this, or, as a matter of fact, of any other country, yet 
gold holds the first place, for it is the easiest to mine and does not 
depend upon a market. 

The three principal districts where gold mining is now being carried 
on are Ambos Camarines, Benguet, and Masbate. 

In the first-named district, near Paracale, mining is largely confined to 
dredging, although vigorous development on the lodes is in progress. The 
country rocks are largely schistose diorites, gneissic granite, and slaty 
shales. These formations are very similar to those found on the Islands 
of the Karimoun Archipelago near Java. The veins, as far as we now 
know them^ are of only moderate width, but very rich; pockets running 
as high as one or two thousand dollars United States currency a ton have 
been encountered. Some of the values are found in the quartz filling 
along the contact of the diorite and granite, but several good veins are 
normal to this contact. Lode mining is farther advanced near Mambulao, 
there being 20 stamps dropping on the San Mauricio property and a 
Huntington mill operating on the Tumbaga. On the Paracale River 
there are two New Zealand dredges, while an American (Risdon) dredge 
is digging in the Malaguit. A large Bucyrus dredge will shortly be 
installed on the Gumaus River. 

The dominant rocks in Benguet are diorite and andesite, and the gold 
is found in fairly large quartz and calcite veins in north and south, east 
and west, and northwest and southwest systems, cutting the formations 
indiscriminately; that is, they are fissure veins. 

Gold occurs both native and as telluride. The ore is not entirely 
free-milling, and cyaniding is resorted to. The chief characteristics of 
this ore, except in one or two cases, are the general absence of manganese 
and the adrupt change from an all-quartz gangue to all-calciie within very 
short distances. 



332 SMITH. 

Altogether, there are only nine stamps dropping at the present time, 
two mills having been badly crippled in the fall of 1909. However, two 
new mills are about to begin operations. 

In the mining district of Masbate the country rock is largely andesite, 
in old, worn-down volcanic stocks. The andesite is cut by a system of 
northwest and southeast quartz veins, one of which is OYeT^^£^metevs wide. 
The ore is oxidized to a much greater extent than is the case in Benguet, 
but is only partially free-milling. This ore also contains large amounts 
of manganese oxide. There are four principal companies in this district, 
but not one in the producing stage at the present time. Some dredging 
was attempted in this district, but for various causes it was unsuccessful. 

Other promising regions are near Cabua and in Nueva Ecija, Luzon ; 
the Cagayan-Muniqui-Pigtao district, Mindanao ; the Surigao Peninsula, 
Mindanao; the Island of Catanduanes; Pangasinan Province, Luzon; 
and Mindoro Island. 

The gold mined in 1909 amounted to 495,194 pesos. 

Silver. — Some silver is found in all the gold districts, usually alloyed 
with gold. Native silver has been encountered in Benguet in small 
quantities. Practically all the lead is argentiferous. The production in 

1908 was 2,750 pesos (1,375 dollars United States currency). That for 

1909 was not e3timated separately. 

Copper. — Copper has been found in the form of arsenates and sulphides 
in the Mancayan-Suyoc district of the Mountain Province, Luzon, as 
native copper in Masbate, the Camarines, and the Island of Jolo. The 
best known deposit is that in the Mancayan-Suyoc district. Eveland ^^ 
writes the following concerning this deposit : 

In view of the fact that the entire region, with the exception of the one ore 
body of the Mancayan mine, is in an early stage of development, it is impracticable 
to treat the ore deposits in detail. It seems to be fairly conclusive, however, that 
the general type of vein in the district is a narrow quartz lead, carrying metallic 
sulphides, in some cases of copper, and generally with gold associated in a free 
state. These veins are in the Mancayaif diorite which underlies the entire 
district. With the advent of the "trachyte" flow, metamorphic changes have 
taken place and the nature of the country rock is altered to a considerable degree. 

Eecent development work has shown this deposit to be rather in the 
nature of a stockwork, and it appears to me to be more extensive than 
was thought to be the case. 

Lead and zinc. — Lead and zinc are associated in some of the Cama- 
rines deposits. Argentiferous lead occurs in many localities, principally 
in Cebu and the Island of Marinduque. In Cebu it is of the nature of 
a stockwork in an andesite flow. 

Iron. — Iron ore in the form of hematite, magnetite, and limonite, 

^Bull P. /., Min. Bur. (1906), No. 4, 53. 



GEOLOGY OF THE PHILIPPINE ISLANDS. 



333 



with varying percentages of iron, has been found in several localities. 
However, there is a fairly well-defined belt which follows, the east coast 
of Luzon for a short distance, beginning in the Camafines and then 
swinging into the interior, where it reaches, as far as our knowledge now 
goes, its greatest development near the town of Angat in Bulacan Prov- 
ince.2^ Three small blast furnaces of crude design are now operated by 
a Filipino woman. 

The following is an analysis of the ore.^^ 

Analysis of the iron ore from Angat, 



Constituent. 



SiOa -. 
AI2O3 
FeO _. 
FeaOa ■ 
MgO_ 
CaO _ 



2.24 
6.52 
1.92 
88.22 
0.18 
0.12 



Constituent. 



NagO 
K2O - 
HoO - 
CO2- 
TiOo - 



Per cent. 



0.04 
0.01 
0.77 



Manganese, etc. — Manganese oxide occurs in Ilocos Norte in nodules 
concentrated in shallow beds. The original location of this substance 
has been traced by Smith ^^ to small stringers in a recent andesite flow. 
Manganese occurs in small pockets in the quartz veins in the gold districts 
of Benguet, Masbate, and Camarines. 

THE NONMETALS. 
FUELS. 

Goal. — Coal is found on nearly every island of the Archipelago. The 
principal localities are Batan Island, Albay Province, Luzon; near Com- 
postela and Danao, Cebu; near Cataingan, Masbate; near Bulalacao, 
Mindoro ; near Sibuguey Bay, Mindanao, and near Burdeos, Polillo. The 
coal measures belong to the Tertiary period. 

The coal is sub-bituminous, and an analysis (by Fox) from Batan 
Island gives the following figures : 2* 9 



Volatile combustible 

Fixed carbon 

Ash 

Moisture 

Sulphur 



39.98 

48.80 

5.70 

5.74 

0.66 



21 H. D. McCaskey, Bull. P. /., Min. Bur. (1903), 2, No. 3. 
22 P. L. Stangl, Ihid. (1903). 

** Asbestos and Manganese Deposits, This Journal^ 8eo. A (1907), 2, 145. 
"Smith, W. D.: The Coal Deposits of Batan Island. Bull. P. /., Min. Bur, 
(1905), 3, No. 5. 



834 SMITH. 

A sample from Gebu gave the following analysis (by Cox) : 

Volatile combustible 36.44 

Fixed carbon 48.80 

Fixed carbon 54.58 

Ash 4.12 

Moisture 4.86 

Sulphur 1.88 

The production in 1909 was 31,000 metric tons. 

There is only one producing mine at present: The East Batan coal 
company on the little Island of Batan. 

Petroleum. — Oil seeps have been noted at two places on the Island 
of Cebu, two in Leyte, four on Tayabas Peninsula, and one in the Cota- 
bato Valley in Mindanao. Three oil wells have been drilled, but no 
appreciable flow has as yet been obtained. One well in Cebu reaches 
a depth of 1,100 feet. The oil occurs in Tertiary sandstone and shales. 

Analysis of oil from Taydhas.^^ 

I. II. 

Specific gravity of filtered oil at 15. °5 C. 0.845 0.831 

Initial boiling point 80°C. 70°C. 

Percentage. 

First fraction, light oils 70° to 150 C. 27.0 36.5 

S'econd fraction, burning oil 150° to 300° C. 56.75 48.75 

Residium above 300° C. by differences 16.25 14.76 



Total 100.00 100.00 

Phosphate, — Small deposits of guano have been found in limestone 
caves, but no rock phosphate. One small deposit of apatite crystals has 
been found in Ilocos Norte. 

Sulphur, — Small deposits, scarcely large enough to be of commercial 
value, occur in Leyte on the Island of Biliran, Mount Apo in Mindanao, 
and a few other scattered localities. Ferguson found a fair deposit on 
the Island of Camiguin in the Babuyanes group. 

Salt, — No large natural salt beds are yet known in the Philippines. 
Some fjry salty springs occur in several provinces, principally in th,e 
Mountain Province of Luzon. 

Ashestos.^ — ^Asbestos of the anthophyllite and fcaltimorite varieties 
occurs in Ilocos Norte, but as yet only occasional veinlets of chrysolite 
have been seen^ There is no production and but little development work. 

Mineral waiers. — The two principal springs are at Sibul, Bulacan, and 
Los Banos, Laguna. Two hundred and sixty-eight thousand four hun- 
dred and forty liters were bottled and sold in 1908. 

... *G. F. bjchmond. Tayabas Petroleum. This Journal, Sec. A (1910), 5, 1. 



GEOLOGY OF THE, PHILIPPINE ISLANDS. 335 

« 

Clay, — Crude ware, pilones, cooking pots, brick and tile, all burning 
red and unglazed, are made throughout the Islands. Some glazed ware 
has been manufactured, using common salt glazing, but little fine faience 
has been attempted. 

In Laguna Province some deposits of kaolin are located, but they are 
not extensively worked. Orthoclase feldspar and sand suitable for glass 
making have been found in a very few places. Good glass sand has 
recentlv been discovered in Ambos Camarines. i 



Although the greatest application of geology is to mining, its relation 
to agriculture is of growing importance and not always appreciated. A 
mechanical analysis of soils is perhaps the most important for the scien- 
tific agriculturist, but next comes that of the chemical composition, and, 
finally, a knowledge of the geologic formation throws much light on the 
interpretation of the results. 

Soils are classified in several ways, namely, according to size of the 
grains, the chemical composition, origin, and lastly for what they are best 
suited. 

The unconsolidated part of the earth's crust, of which the soil is only 
the uppermost layer, is known as the regolith or blanket rock. According 
to Merrill ^^ the subdivision of the regolith are as follows : 

fR 'd 1 (Residuary gravels, sands and clays, 

Sedimentary— J [ wacke, laterite, terra rossa, etc. 

jcumulose.... Peat, muck, and swamp soils, in part. 



The regolith,.. 



P ,, . •, [Talus and cliff debris, material of 
I avalanches. 

{Modern alluvium, marsh and swamp 
(paludal) deposits, the Champlain 
clays, loess, and adobe, in part. 
iEolian (Wind-blown material, sand duneS, 

I adobe and loess, in part. 
Glacial (Morainal material, drumlins, eskars, 

I osars, etc. 

Leaving out of account the relative merits of the systems mentioned 
on the previous page, I shall discuss the soils of the Philippines in relation 
to their origin. This course seems best for two reasons : First, it is the 
only one which a geologist should attempt to take, and, second, it is my 
own belief that the only true classification is a genetic one. 

Two great subdivisions might then be made as follows : 

^ A Treatise on Rocks^, Rock- Weathering and Soils. New York, The Mab!- 
millan Company (1^06), 228. 



336 



SMITH. 

PHUppine soils. 



Soils. 


Origin. 


Resulting soil. 


Residual 

Transport;pd__ 


Basalt 

Andeslte 


JLaterite. 

Clay with much insoluble matter, poor soil. 

Clay loam. 

Sandy loam to very poor sandy soil. 

Silt. 

Sand, gravel, etc. 


Limestone 

Shale 


Sandstone 

j Fluvial, 


Subaerial 





As the rocks in the Philippines are largely volcanic and as the pre- 
dominating class is andesite^ or the olivine variety of it which is termed 
basalt, it is to be expected that a large area of the Islands would be 
covered by a soil derived from the disintegration of these rocks, and such 
is the case. This soil is usually heavy, red, and because of the large 
amount of magnesia, iron, calcium, alumina, etc., which it contains, it is 
exceedingly fertile. This ferruginous soil so much resembles some of 
the material in India which has been called laterite that I shall so term it 
in this discussion. However, all of the formations classed in India as 
laterite have not originated in the manner outlined in the beginning 
of this paragraph. 

Examples of this soil are to be found covering large tracts of Bataan 
Province, Luzon ; much of western Mindanao and its upland, the vicinity 
of Sarangani Bay on the same island ; Negros ; northern Luzon, especially 
near Baguio; Basilan and Albay, Luzon. This soil is much like that of 
the Hawaiian Islands and is very well adapted to sugar growing, rubber, 
hemp, etc. 

Limestone is the next formation which plays an important part in the 
geology of the Islands. Vast areas are covered by this material and it is 
found on the summits of some of the highest mountains, such as Mount 
Data in Luzon. The formation is largely derived from triturated coral. 
Although a limestone largely is made up of soluble matter, it almost 
always contains some insoluble clay; the latter stays behind while the 
soluble part is carried off by streams. Therefore, a limestone formation, 
contrary to the usual belief, does not- yield a very fertile residual soil. 
On the other hand, the transported soil originating from limestone is 
exceptionally fertile. As an instance, we need only point to the limestone 
covered tracts of parts of Cebu. There the crops, chiefly corn, are very 
poor, while in the coves some of the best agricultural land in the Islands is 
to be found. 

The coal measures, which include sandstone shale and clays, are 
productive of very good soils. Sandstone ordinarily is too high in un- 
combined silica to afford plants much nourishment, but the saadstone in 



GEOLOGY OP THE PHILIPPINE ISLANDS. 337 

the Philippines has been derived from the disintegration of rocka poor 
in quartz. It is more in the nature of an arkose, and hence yields a better 
soil than is usual for sandstone. The shales of the coal measures, next 
to the volcanic rocks, I believe, yield what may be classed as among the 
best residual soils. 

In temperate regions we must consider several varieties not found in 
the Tropics ; for instance, glacial soils and loess or wind-borne soils, but 
in the present discussion concerning the soils of the Philippines only the 
water-borne material, fluvial, and subaerial soils remain for consideration. 

Examples of fluvial soils are seen in the valley of the Cagayan in 
northern Luzon, or in the Cotabato Valley, Mindanao. These soils are 
very rich, not only because of their chemical heterogeneity, but also 
because of the mechanical state of their particles, they being in an un- 
usually finely divided condition and hence more easily utilized by plants. 
Furthermore, in a broad meandering stream the soil is being continually 
worked over by the lateral cutting and filling familiar to one who has 
much acquaintance with rivers. 

The subaerial deposits are water-borne, with the difference that the 
material is not carried along definite channels, but is spread out sheet- 
like in fans and cones at the bases of the mountains. The material is 
always much coarser than that just refered to and therefore not so good. 

A heavy, sticky, grayish-blue soil, which corresponds pretty closely to 
the ^^adobe'^ in California, occurs at the bases of many mountains in the 
Islands. This is derived largely from the decomposition of basalts and 
similar rocks. This soil, because of its impervious character, is especially 
good for rice growing, as it holds water for a long time. 

From my own observations, while in the field working on other matters 
more strictly geologic, I have noted that the relation between the agri- 
cultural products and the geologic formations in general is shown in the 
following table : 



Residual- 



Transported 



Laterite Sugar, coffee, camotesi" some hemp, coco- 
nuts. 

Limestone - Com, camotes. 

Sandstone and shales... Hemp. 

Fluvial Tobacco. 

Subaerial Hemp and some sugar. 



As I have made no special studies along this line, this list should be 
regarded as only a rough tentative summary. It is hoped that it will be 
suggestive to scientific agriculturists. 

Artesian water. — The principles governing the occurrence of artesian 
water are so generally known that they need not be repeated. Still, it is 

" Sweet potatoes. 



388 SMITH. 

not an uncommon thing for wells to be sunk apparently without the 
slightest regard to the geology of the district wherein the well is located. 
I know of several cases like this, one where a well site was selected by 
means of the divining rod and later had to be abandoned ; a second well 
was drilled for twenty-two months in the basal igneous complex ; a third 
tapped a stratum which outcropped a short distance away in salt water. 
However, such mistakes are not characteristic of the Philippines, for the 
Bureau of Public Works has been particularly successful in its well work. 
Although this organization has drilled wells in many parts of the 
Archipelago, it has met with its greatest success in the central plain of 
Luzon where artesian conditions are most favorable; that is, where 
alternating tuff and sandy strata with intercalated layers of clay afford 
good water passages. Many of. these wells do not exceed 100 meters in 
depth and yield a good supply of potable water. 

Some of the conditions favorable for the sinking of successful wells may 
be enumerated as follows : 

1. A porous stratum (preferably a sandstone) with impervious strata 
(clay) above and below, flexed into a syncline or monocline and out- 
cropping in a more elevated and rainy region. 

2. Porous volcanic material, ash or tuff, overlain by impervious 
material. 

3. Porous volcanic material near a coast fringed with coral reefs 
(ponding). 

The conditions are best fulfilled in the piedmont areas where moder- 
ately deep wells tap the coal measure sandstone lying on the flanks of the 
Cordilleras, or in the Manila tuff formation. 

LIST OF THE MORE IMPORTANT WORKS ON PHILIPPINE GEOLOGY.-^ 

Abella y Casabieqo, Enrique. Memoria acerca de los criaderos aurlferos del 

segundo departamento de Mindanao, Misamis. Bol. de la Com. del Mapa 

geol. de Espana, Madrid, Tello (1879), 8o, 49 pp., 5 pis. 
Idem. Itinerarios geologicos. Observaciones tomadas al paso en los viajes he- 

chos A las comarc^as aurfferas de Misamis. Bol. de la Com. del Mapa geol. 

de Espana (1879), No. 6, 65-81, 1 map. 
Idem. Terremotos de Nueva Vizcaya (Filipinas) en 1881, etc. Madrid, Tello 

(1884), 80, 31 pp., 1 map. 
Idem. La Isla de Biliran (Filipinas) y sus azuf rales. Madrid, Tello (1884), 

8o, 11 pp., 1 pi. (map). Also in Bol. de la Com. dil Mapa geol. de Espana 

(1884), No. 11. 
Idem. El Mayon 6 Volcan de Albay (Filipinas). Madrid, Tello (1885), 8o, 

23 pp., 2 pis. Also in Bol. de la Com. del Mapa geol. de Espana (1884), 

No. 11. 

, ^ These are tak^n from a complete bibliography being prepared by Mr. Henry 
G. Ferguson. 



GEOLOGY OF THE PHILIPPINE ISLANDS. 339 

Idem. El monte Maquilin (Filipinas) y sus actuales emanaciones volcftnicas. 

Madrid, Tello (1885), 8©, 28 pp., 2 pis. Also in Bol. de la Com, del Mapa 

geol. de Espana (1884), No. 11. 
Idem. R^pida descripci6n fisica, geol6gica y minera de la Isla de Cebti. Madrid, 

Tello (1886), 187 pp., 5 pis. 1 map. Also in Bol. de la Com. del Mapa geol. 

de Espana (1886), No. 13. 
Idem. Descripcion ffsica, geologica y minera de la Isla de Panay, Manila, Chofr6 

(1890), 80, 203 pp., 2 maps, 1 table. 
Adams, George I. The Marble and Schist Formations of Romblon Island. This 

Journal, Sec. A (1909), 4, 87-89. 
Idem. Geological Reconnaissance of Southwestern Luzon. This Journal, Sec. A 

(1910), 5, 57-117. 
Idem. Geological Reconnaissance of the Island of Leyte. This Journal, Sec. A 

(1909), 4, 339. 
Idem. Nonmetallic Mineral Products. Production of Structural Material. Min. 

Res. P. I., Bureau of Science, Manila (1909), 45-49. 
Bacon, Raymond F. The Crater Lakes of Taal Volcano. This Journal, Sec. A 

(1907), 2, 115. 
Becker, George F. Report on the Geology of the Philippine Islands. 21st Ann. 

Rep. U. S. Geol Surv. (1901), 3, 493-614, 2 maps, 1 pi. 
Burritt, Charles H. Report of Chief of Mining Bureau, P. I., for year ending 

May 31, 1901. First Annual Report of the Philippine Commission. Washing- 
ton (1901), Appendix K. K. 
Idem. Report of the Chief of the Mining Bureau for the year ending September 

1, 1903. From the Fourth Annual Report of the Philippine Commission. 

Washington (1903), 325-342. 
Centeno y Garcia, Jost. Ministerio de Ultramar. Memoria geol6gica-minera 

de las Islas Filipinas. Madrid, Tello (1876), 8o, 8, 64 pp., 1 map. This 

appeared separately and also in Bol. de la Com. del Mapa geol. de Espana 

(1876), No. 3, 183-234. 
Cox, A. J. The Proximate Analysis of Philippine Coals. This Journal, Sec. A 

(1907), 2, 41-65. 
Idem. Philippine Coals as Fuel. This Journal, Sec. A (1908), 3, 301-356. 
Idem. Philippine Raw Cement Materials. This Journal, Seo. A (1909), 4, 

211-229. 
Brasche, Richard von. Ausfluge in die Vulcangebiete der Umgegend von Ma- 
nila. In Verhandlungen der K.-K. geologischen Reichsanstalt (1876), 89-93. 
Idem. Meittheilungen aus den Philippinen. In Verhandlungen der K.-K. geolog- 
ischen Reichsanstalt (1876), 193-198. 
Idem. Aus dem Siiden von Luzon. In Verhandlungen der K.-K. geologischen 

Reichsanstalt ( 1876 ) , ' 251-255. 
Idem. Einige Worte iiber den geologischen Bau von Siid-Luzon. In Mineralog- 

ische Mittheilungen gesammelt von Gustav Tschermak, Wien, HSlder (1876), 

157-166. 
Idem. Fragmente zu einer Geologie der Insel Luzon (Philippinen), etc. Vienna, 

Gerold's Sohn, (1878), 99 pp., 5 pis. (incl. 2 maps). 
Eddingfield, F. T. The Baguio Mineral District. Min. Res. P. /., Bureau of 

Science, Manila (1910). 
Espina y Capo, Lins. Ligero bosquejo acerca de los principales yacimientos me- 

tallferos de Filipinas, Manila. MS. (1898), 191 pp. 
99514 3 



340 SMITH. 

EvELAND, A. J. A Preliminary Reconnaissance of the Mancayan-Suyoc Mineral 

Region, Lepanto, P. I. Bull P. I, Min. Bur., Manila (1905), 2, No. 4, 58 

pp., 43 pis., 6 maps. 
Idem. Notes of the Geology and Geography of the Baguio Mineral District. 

This Journal, Sec. A (1907), 2, 207-233. 
Ferguson, H. G. A Geological Reconnaissance of the Batanes Islands. Far 

Eastern Review (1907), 4, 152-153. 
Idem. The Sulphur Deposits of Camiguin Island. Faf Eastern Review (1907), 

4, 152-153. 
Idem. Contributions to the Physiography of the Philippine Islands. TI. Bata- 
nes Islands. TUs Journal, Sec. A (1908), 3, 1-25, 3 maps, 9 pis. 
Idem. Physiography of the Philippine Islands, III. Western Masbate. This 

Journal, Sec. A (1909), 4, 1-17, 3 maps. 
Idem. The Metallic Resources of the Philippines. Min. Res. P. I., Bureau of 

Science, Manila (1909), 20-29. 
Idem. The Aroroy District, Masbate. Min. Res. P. I., Bureau of Science, Manila 

(1910). 
Goodman, Maurice. Island of Marinduque. Far Eastern Review (1907), 3, 

365-366. 
Idem. Gold Placers of Nueva Ecija. Far Eastern Review (1907), 4, 88-89. 
Idem. Sulphur in the Philippines. Far Eastern Review (1907), 4, 120-121. 
Idem. Metallic Mineral Resources. Min. Res. P. L, Bureau of Science, Manila 

(1908), 22-25. 
Idem. A Reconnaissance from Davao, Mindanao, over the divide of the Sahug 

River, including a survey from Davao to Mati. Narrative of the Expedition. 

This Journal, Sec. A (1908), 3, 501-511, 2 maps. 
Idem. The Gold Field of the Surigao Peninsula, Mindanao. Min. Res. P. I., 

Bureau of Science, Manila (1909), 40-44. 
HocHSTETTER, F. VON. Schreiben an Alexander von Humboldt. Sitzungsber. 

Akad. d. Wiss. math.-nat. Klasse, Wien (1859), 121-142, 1 map. 
ICKis, H. M. Polillo Coal Deposits. 19 pages MS. in Bureau of Science, 1 map 

(1906). 
Idem. Camarines Gold Fields. Far Eastern Review (1907), 4, 56-57. 
IDDINGS, J. P. The Petrography of Some Igneous Rocks of the Philippines. This 

Journal, Sec. A (1910), 5, 155. 
Jagor, F. Reisen "in den Philippinen. Berlin, Weidmannsche Buchhandlung 

(1873), 16, 381 pp., 1 map. 
Koto, B. On the Geologic Structure of the Malayan Archipelago. Journ. Coll. 

Sci. Imp. Univ. Tokyo (1890), 2, pt. 2, 83-120, 1 map. 
McCaskey, H. D. Report on a Geological Reconnaissance of the Iron Region of 

Angat, Bulacan. Bull. P. I. Min. Bur., Manila (1903), No. 3, 62 pp., 42 pis., 

6 tables, 1 map. 
Idem. Fifth Annual Report of the Mining Bureau for the year ending August 

31, 1904. Manila (1905), 44 pp., 13 pis., 2 maps. 
Idem. Sixth Annual Report of the Chief of the Mining Bureau, Manila (1905), 

66 pp., 14 pis., 2 maps. 
Idem. First Annual Report of the Chief of the Division of Mines to the Director 

of the Bureau of Science, being for the period from December 1, 1905, to 

July 1, 1906. MS. 52 pp., Manila (1906). 



GEOLOGY OF THE PHILIPPINE ISLANDS. 341 

Mabche, Alfred. Lugon et Palaouan, six ann^es de voyage aux Philippines. 

Paris, Hachette (1887), 398 pp. 
Martin, K. Ueber tertiare Fossilien von den Philippines. Samn. des geol. 

Reichs-Museums in Leiden (1896), 5, 52-69. 
Mas6, M. S. Volcanoes and Seismic Centers of the Philippine Archipelago. 

Census of Phil, Islands Bull., Washington (1905), 3, 80 pp., 4 maps, 5 pis. 
Idem. The Camiguin Volcano. Bull. P. I. Weather Bur., Manila (1902), 162- 

129, 1 pi. 
MoNTANO, J. Rapport Sl M. le ministre de Tinstruction publique sur une mission 

aux iles Philippines et en Malaisie (1879-1881). (Paris Hachette 1886.) 
Moses, A. J. The Crystallization of Luzonite and Other Crystallographic Studies. 

Amer. Journ. 8ci. (1905), IV, 20, 277-280. 
Nichols, J. C. Notes on the pigholugan and Pigtao Gold Region, Island of 

Mindanao. Trans. Amer. Inst. Min. Eng. (1901), 31, 611-616. 
Oebekke, K. Beitrage sur Petrographie der Philippinen und der Palau-Inseln. 

Neues Jahrb. f. Mineral, etc., Beilage, Band I, Stuttgart Koch (1881), 

461-501. 
RiCHTHOFEN, F. VON. Ucber d|is Vorkommen von Nummuliten Formation auf 

Japan und den Philippinen. Ztschr. d. deutschen Oes. (1862), 357. 
Roth, Justus. Ueber die geologische Beschaffenheit der Philippinen. Jagor, F., 

Reisen (1873), 333-354. 
Idem. Construccion geol6gica de Filipinas. In Jagor's Viajes por Filipinas. 

Madrid (1875), 349-373.^ 
Sainz de Baranda. -Constitucion geognostica de las islas Filipinas. Ann. Minos, 

Madrid (1841), 2, 197-212. 
Santos, Jos^ Maria. Informe sobre las minas de cobre de las rancherfas de 

Mancayan, Suyok, Bumucun y Agbao en el Distrito de Lepanto, etc. Manila, 

Press of the College of Santo Tomfts (1862), 72 pp. Also in Revista minera, 

Madrid, 14. 
Semper, Carl. Reise durch die nordostlichen Provinzen der Insel Luzon. Ztschr. 

f. allg. Erdkunde, 7ieue Folge, Berlin, Reimer (1861), 10, 249-266. 
Idem. Reise durch die nordlichen Pronvinzen der Inseln Luzon. Ztschr. f. allg. 

Erdkunde, neue Folge, Berlin, Reimer (1862), 13, 80-96. 
Smith, W. D. The Coal Deposits of Batan Island. Bull. P. I. Min. Bur., Manila 

(1905), No. 5, ^Q pp., 11 photographs, 11 maps and sections. 
Idem. Orbitoides from the Binangonan Limestone. This Journal (1906), 1, 

203-209, 2 pis. 
Idem. Preliminary Geological Reconnaissance of the Loboo Mountains of Ba- 

tangas Province.' This Journal (1906) 1, 617-631, 4 pis. 
Idem. Contributions to the Physiography of the Philippine Islands. Cebti Island. 

TUs Journal (1906), 1, 1043-1056, 8 pis. 
Idem. Petrography of Some Rocks from Benguet Province, Luzon, P. I. This 

Journal, Sec. A (1907), 2, 235-253, 5 pis. 
Idem. The Geology of the Compostela-Danao Coal Field. This Journal, 8eo. A 

(1907), 2, 377-403, 15 pis., 3 maps. 
Idem. The Nonmetallic Minerals.. Min. Res. P. /., Bureau of Science, Manila 

(1908), 11-21. 
Idem. The Asbestos and Manganese Deposits of Ilocos Norte, etc. This Journal, 

Sec. A (1907), 2, 145. ^ 



342 SMITH. 

Idem. A Geologic Reconnaissance of the Island of Mindanao and the Sulu 
Archipelago. I. Narrative of the Expedition. This Journal Sec. A (1908), 
4, 473-500, 23 pis. 

Idem. A Geologic Reconnaissance of the Island of Mindanao and Sulu Archipelago. 

II. Physiography. This Journal, Sec. A (1910), 5. 

Idem. A Geologic Reconnaissance of the Island of Mindanao and Sulu Archipelago. 

III. Geology. This Journal, Sec. A (1911), 1. 

Idem. The Nonmetallic Minerals. Min. Res. P. L, Bureau of Science, Manila 

(1909), 11-19. 
Idem. Coal Resources of the Philippines. Economic Oeology (1909). 
Idem. The Paracale-Mambulao Gold District. Min. Res. P. /., Bureau of Science, 

Manila (1910), 1. 
Idem. The Philippine Islands. Handbuch der Regionalen Geologic. Herausge- 

geben von Steinmann u. Wilkens Heidelberg (1910), Bd. VI 5, Htft 3. 



ILLUSTRATIONS. 



Plate I. Pagsanjan Falls. 

II. Map. Showing orography and hydrography. 

III. Map. Showing general geology as known at the present time. 

IV. Characteristic Philippine fossils. ' 
V. Map. Showing principal mining districts. 

VI. View of Gold Creek, Benguet, and surrounding country. 

TEXT FIGURES. 

Fig. 1. Section across northern Luzon. 

2. Section across Luzon (vicinity of Manila). 

3. Section across a part of Cebu. 

4. Section across a part of Mindanao. 

343 



Smith : Geology of the Philippine Islands.] [Phil. Joubn. Sci., Vol. V, No. 5. 




RLAXE I 



• • »•. ••• 



Smith : Geology of the Philippine Islands.] [Phil. Journ. Sci., Vol. V, No. 5. 



4^ 



MAP SHOWING RELATION I '•^'''''^ 

^_ ol • Active ordormanr voice-^oes 



RECENT VULCANISM 

TO Babuyan Ciarolj 



"♦- Extinct or soirateric volcanoes 



Tectc^ic lines 
a Inner series of Becker 
r^r^.v.^ ''^^"^^'■"It D Ourer series of Becker 

PRINCIPAL TECTONIC LINE.^- L c Sierra Madre l.ne 




RLATE II. 



Smith : Geology of the Philippine Islands.] 



[Phil. Journ. Sci., Vol. V, No. 5. 



-T- 



7i7" 



«> P 




Tuff and 
agglomerates. 



Tertiary 
sedimentaries. 



S 



Extrasive Complex of igneoas and Metamorphlo 

rocks. metamorphio rooks. rooks. 



RLATE III. 



Smith : Qeoloqy of the Philippine Islands.] 



[Phil. Journ. Scr., Vol. V, No. 5. 




1. Vicarya callosa Jenk var. 

semperi Sm. 

2. Vicarya callosa Jenk. 

3. Rostellarea javana Mart. 

4. Portion of cast of Vi- 

carya. 

5. Turricula bataviana 

Mart. 

6. Natica globosa Chem. 



7. Natica caniculata Lam. 

8. Hindsia dijki Mart. 

9. Oliva (ancilla) sp. nov. 

10. Cardita boettgeri Mart. 

11. Venus chlorotlca Phil. 

12. Cytherea allied to ventrl- 

cola. 

13. Cardium rugosum Lam. 



14. Lucina (Codakia) allied 

to semperiana Issel. 

15. CultellusmaxlmuB. Qme- 

lin. 

16. Odontocyathus sp. nov. 

17. Plabellum australe (?). 

18. Pachyseris a r i s t a t a 

Mart. 



PLATE 



IV. 



Smith : Geology of the Philippine Islands.] 



[Phil. Journ. Scl, Vol. V, No. 5, 




RLATE V. 



f- 'i'' 



GEOLOGIC RECONNAISSANCE OF MINDANAO AND SULU: 
.II. PHYSIOGRAPHY. 



By Warren D. Smith. 
(From the Division of Mines, Bureau of Science, Manila, P. I.) 



COXTENTS. 
Physiography. 

A. Plains. 

B. Rivers. 

* C. Intermediate uplands. 

D. Cordilleras. 

E. Minor types. 

1. Lakes. 

2. Coral reefs. 

3. Terraces. 

4. Volcanoes. 

5. Atolls, crater lakes, etc. 

In taking np the second chapter in the Reconnaissance of Mindanao and 
Sulu, I am fully aware of the limited amount of our present knowledge 
of that vast territory. Much of this information is scattered and can be 
found only in old libraries. For instance, one very important contribution 
by Abella ^ on the Misamis placer fields I am sure has never been seen 
by many English-speaking people for the reason that it is published in 
a bulletin of the Commission of the Geological Map of Spain. Similarly, 
important notes on the Agusan Eiver region are to be found as a part of 
a report made by M. Montano to the Department of Public Instruction 
of France.- The observations of our own party are also necessarily very 
fragmentary owing to the very unsettled state of the country. In the 
following paper I shall discuss the territory in question, from a physi- 
ographic point of view solely.. Part III will take up the geology and 
mineral resources. 

^Abella y Casariego, E. Memoria acerca de los Criaderos Auriferos del Se- 
gundo Distrito del Departamento de Mindanao. Bol. Del. Mapa Oeol. de Espo/fla 
(1877), 6. 

* Montano, J. Une Mission aux les Isles Philippines. Rapport a M. Le 
Ministers de L'Instruction Publique de France (1879-1881). 

345 



346 SMITH. 

The great Island of Mindanao extends from nearly 6° 30' to 10° north 
latitude and from 122° 30' to 126° east longitude, a region of lofty mount- 
ains, beautiful lakes, primeval forests, and long and deep rivers. The 
European settlements, all on the coast, can be counted on the fingers of 
one hand. Two military stations are located in the interior, one of 
considerable size, the other a mere outpost. Army transports and a few 
trading boats stop at some points to discharge cargo and land new relays 
of troops. With the exception of the soldiers, those who visit the island 
rarely go more than 8 to 16 kilometers inland, so that as far as the outside 
world of commerce, tourists, and even scientists is concerned the interior, 
with all its strange people^ its latent wealth, and its magnificent scenery, 
is but little known. 

PLAINS. 

Mindanao has not a great proportion of coastal plain. Sufficient 
time has not elapsed for its formation to any great extent, as mountains 
everywhere border the sea. The only coastal plain of any importance 
is that of Zamboanga ;* it is confined to the lower end of the peninsula 
of the same name and roughly is 12 by 35 kilometers in area. It has a 
foundation of coral reef with a veneer of piedmont deposits above it. One 
small stream, the Tumaga, straggles across it to the sea. The plain is 
quite fertile and is the home of a lafge and mixed population of Moros, 
Filipinos, Chinese, Germans, and Americans. The city of Zamboanga, 
because of the fertility of the sorrounding plain and its exceptional situa- 
tion, occupying a ^^gate^^ position with respect to southern Mindanao and 
the Celebes, has risen to the first rank among the settlements of the south. 
However, there is one factor which will always work against its greatest 
development: the Tumaga Eiver is not navigable into the back country, 
which, furthermore, is exceedingly rugged and not capable of supporting 
other than an uncertain population, either of restless hill tribes as at 
present, or of mining camps which have not yet developed. 

The only comparatively large plains in Mindanao border the Cotabato 
and the Agusan Eivers, which are approximately 400 and 300 kilometers 
in length. The width of these valleys varies from only a few kilometers 
to 50 or 60. The material constituting these river plains is for the 
most part fine alluvium derived from many classes of rocks passed over 
by the rivers en route. It is, of course, most excellent soil. 

THE COTABATO PLAIN. 

Beginning at the lower end of the Eio Grande de Mindanao (Cota- 
bato), a true delta formation is found, the river making its way across 
this through several channels; two much larger than the others, and 
debouching in four different places. The two main branches are the 
Cotabato (north), and the Tamontaca (south). Between these a net- 



MINDANAO AND SULU : II. PHYSIOGRAPHY. 347 

work of estuaries affords easy communication by boats between the many 
settlements on the plain. The true delta runs back as far as the junction 
of the two main branches at Tumbao. The valley above this, for 30 
kilometers or so, is not so wide, because the low Silik hills come down 
fairly close to the river on the north side ; the latter then makes almost a 
right-angled bend to the north. To the southeast of this point there is 
a wide, low tract, 50 or 60 kilometers across, containing two bodies 
of water, marked on the map as Lakes Liguasan and Buluan. These 
are little more than swamps, their size varying with the rainfall. This 
great tract is gradually draining as the land slowly rises.^ However^ 
it could be artificially drained should this island ever become extensively 
colonized. 

From this point to the northward the valley plain continues with 
varying width. Some low hills appear to the northwest at the junction 
of the Kabakan and the Kio Grande, but toward the sea there is an 
extensive tract of low, forest-covered land without the high mountain 
range usually shown on the old maps. This position, which is about 160 
kilometers from the mouth of the river, is the farthest locality I reached 
in traveling up the valley. From this, point I turned off to the east, 
following up the Kabakan Eiver. Mr. Ickis, who made the trip from 
Cagayan, in Misamis, to Sevilla, reported extensive open country on either 
side of the river in that vicinity. 

It seems reasonably certain that the region to the south of the Rio 
Grande was once separated from the northern part by an arm of the sea 
which extended from Cotabato to Sarangani Bay. The existence of raised 
coral reefs on the south of the river points to this with a fair degree 
of conclusiveness. The stretch of country east of Lake Liguasan is 
underlain by loose sandstone which probably is quite recent. The rocks 
of the Matutan Range, a very interrupted line of volcanic stocks, are 
andesites which have poured over the country in rather recent times. 

The largest settlements of the Maguindanao Moros are to be found at 
the lower end of the Grande River plain. Cotabato, a town of about 
1,000 civilized * inhabitants is situated about 10 kilometers, by river, 
from the mouth. Although Cotabato is behind both Zamboanga and 
Davao at the present time, it has in some respects a more favorable 
geographic situation than either. Any town situated at the mouth of a 
navigable river draining an interior like that of central Mindanao pos- 
sesses great advantages over those which are not so situated; and, there- 
fore, if Mindanao is ever opened to colonizers, Cotabato should become 
the metropolis. 

' The evidences for very recent and very considerable elevation in Mindanao 
are abundant. 

* Philippine Census (1903). 
99514 4 



348 SMITH. 

The Eio Grande de Mindanao ° has its source in the mountains east of 
Cagayan, Misamis, and flows almost due south to the southern boundary 
of the subprovince of Bukidnon, swings west to the town of Sevilla, then 
east again in a wide curve and then southwest; a few kilometers north 
of Lake Liguasan it turns at right angles and flows a few degrees north 
of east into Illana Bay. It is quite probable that the former mouth was 
very close to Lake Liguasan and that the river has gradually grown 
southward as it built up the delta mentioned on page 347. Meanders 
are quite common in this river (see Plate II), there being one point near 
Tumbao where, by taking a small canal of 50 or 75 meters in length, over 
half an hour of travel can be saved. The banks show no rock exposures, 
nothing but mud could be seen as far up as I followed it. The fall is 
only about one in 5,000, as the 30-meter contour does not cross it until a 
point 10 or 12 kilometers above the confluence of the Kabakan and 
Pulangui is reached, or about 160 kilometers from the sea. 

It is difficult to navigate any but light, flat-bottomed craft above the 
Kabakan. A flat-bottomed, stem-wheeled steamship drawing not over 
2 feet should be able to ascend 80 kilometers above this point, and it 
might reach Sevilla, were it not for Murphy's rapids, which Captain 
Murphy describes as occurring in the vicinity of the Alanan Eiver. I 
suspect that an outcrop of some hard, igneous rocks, like diorite, occurs 
here. Unfortunately, the Army topographers, to my knowledge, collected 
no geologic specimens in the course of their work in this portion of the 
Cotabato Valley, and perhaps with good reason, for all through Minda- 
nao traveling and packing is extremely difficult. The United States 
Army engineers have made an excellent topographic map of a large part 
of the Bio Grande and the region it drains, particularly that to the 
north. The topography by Captain Murphy and his associates in this 
region shows that the old maps were very broadly generalized, to say the 
least, and in many instances totally wrong. There is nothing resembling 
the regular, linear arrangement of north and south tributaries as is shown 
on the Spanish maps or the later American editions based on them. 

Terraces along this river are well shown, following the south fork below 
Tamontaca and on the north branch back of Ungup, and again near the 
confluence of the Libugan and the Eio Grande. Between the town of 
Cotabato and old Fort Tamontaca there is an extensive, raised, flat- 
topped platform which rises like an island in the delta. This was 
examined and found to be a raised coral reef 5 to 10 meters above the 
letel of the river. 

The Eio Grande, or Pulangi, overflows its banks periodically and thus 
irrigates and enriches ,an already marvelously fertile soil. Eice is the 
chief crop produced, but the methods of agriculture are very primitive. 

"Also termed the Pulangui and the Cotabato. 



MINDANAO AND SULU: II. PHYSIOGRAPHY. 349 

THE VALLETT OF THE AGUSAN. 

Mr. Maurice Goodman^ to some extent has already described this 
great intermontane flat. However, he does not give aay definite figures 
as to its width. It is very wide, as is attested by the fact that the distant 
mountain peaks could be seen with difficulty. This broad valley must 
be in the neighborhood of 300 kilometers in length. It is inhabitated 
by Christian Yisayans, Manobos, Mandayas, Manguangans, and Bukid- 
nons, the first named having immigrated there. 

Some so-called lakes are found about midway in the length of this 
great stream, but according to all travelers in this region they axe to be 
regarded as large, low, swampy areas over which the river has spread. 
The lakes are supposed to have originated in a local subsidence at the time 
of the earthquake of 1892, which was very much like that causing the 
formation of Reelfoot Lake in western Tennessee in the United States. 
The description given me by an American civil engineer, formerly super- 
visor of Surigao Province, confirms this view. He said that many trees 
of a kind not usually growing in water could be seen submerged so that 
only the top branches were visible. This is not at all unreasonable in 
view of the fact that the valley of the Agusan is a focus of great seismic 
activity, and the line of equal magnetic intensities follows its trend.*^ 

There is little transportation over this plain. The river with its 
tributaries is at present practically the onl} route of travel. Hemp and 
rice are the chief agricultural products. 

SAIALL RIVERS. 

Besides the two main rivers of central and eastern Mindanao and the 
Agusan, in the northeastern district, there are several shorter and un- 
navigable ones in various parts of the island. The following will be 
considered: (1) The Sahug-Tagum system, (2) the Agus, (3) the 
Mataling, (4) the Cagayan, (5) the Iponan, (6) the Tumaga. Others, 
which are much longer than the last mentioned, but as yet of little 
commercial importance, have been omitted. 

(1) The Sahug-Tagum system, — This system, which is known as the 
Sahug in the upper waters and the Tagum in the lower part, has its 
source on the southern slopes of Mount Kuanabayan. This is the route 
Ickis and Goodman took in the early part of 1908 to pass from the Gulf 
of Davao to the upper waters of the Agusan. A much shorter way, 

•A Reconnaissance from Davao/ Mindanao, over thie Divide of the Sahug River 
to Butuan, including a survey fron^ Davao to Mati. This Journal, Sec. A (1908), 
3i 601. 

^Mr. Goodman found that abnormally high water was sufficient to explain 
these lakes. It is quite probable that in 1902, when ihe civil engineer • visited 
them, there was a flood. M. Saderi^a Mas6, S. J. Isoclinic and Isogonic lineii 
in the Island of Mindanao, Phil. Weather Bureau, Manila (1902), p. 246. 



350 SMITH. 

and one more frequently traveled, is by way of the Hi jo to Compostela. 
Practically the only published data regarding the Tagum Eiver are taken 
from Mr. Goodman's narrative. He describes it as a tortuous stream, 
about 90 meters wide at its mouth, but only about 30 at the confluence 
of the Sahug and the Tagum; at this point its banks are 5 meters 
above the water level and the formation is a brown and blue clay overly- 
ing sandstone. At Matinlud the river shoaled so much that a loaded 
hanca could not be floated. The coimtry it traverses is all heavily wooded 
and occupied by Mandayas. Dr. J. Montano, a French traveler, also 
made this trip, but gives us very meager notes on the physiography of 
the region. A map is given in his book showing a long ridge at right 
angles to the x\po Eange and curving to the northwest. Goodman, also, 
mentions this ridge. Montano likewise gives some geologic notes, to which 
I shall refer in the next section of the paper. He published a sketch 
map which is taken probably entirely from d'Almonte. 

(2) The Agus River, — This river drains Lake Lanao and flows north 
to empty into Iligan Bay. It is approximately 32 kilometers long and 
drops from 671 meters to sea-level. It has an average fall of 20 meters 
per kilometer (110 feet per mile). However, this is not uniform, for 
at Maria Christina Falls there is a sheer drop of 58 meters. The river 
is very narrow and swift, flowing over a basalt formation throughout its 
length. The army engineers have ascertained the depth and width at 
Mumungan to be 26.5 and 15.5 meters respectively. The flow of water 
has recently been measured by Mr. Bradshaw of the Bureau of Public 
Works, at Pantar bridge. Here the flow amounted to 9 second-meters, 
giving approximately 40,000 electric horsepower. 

(3) The Mataling is another short river with a considerable fall. It 
rises in tjie ridge south of Lake Lanao, and flows southwest into Illana 
Bay. It likewise flows over a basalt formation, and at Mataling Falls 
drops over a hard layer of this rock into a pool 16 meters below. From 
here on the rock is much softer, and the river a mile or two beyond flows 
through a low plain of loose, bluish-black volcanic ash. The Mataling is 
very narrow, 10 meters on an average, and very swift. Its source in part 
is undoubtedly Lake Lanao, although there is no connection apparent on 
the surface ; it also receives the run-off from Davao Lake, a small body of 
water near the southwest corner of Lake Lanao. 

A very important series of rivers in the northern part of Mindanao 
rises in the mountainous interior of the subprovince of Bukidnon and 
flows north through the Province of Misamis into Makahalar Bay. The 
most important of these are the Iponan, the Cagayan, and the Tagoloan. 
They for many decades, perhaps for centuries, have been favorite localities 
for gold-panning. Ickis traversed and mapped one of them, the Tago- 
loan, throughout its whole length. This river shows some interesting 
features. It rises near Tibua Mountain and is separated by a very 



MINDANAO AND SULU: II. PHYSIOGRAPHY. 



351 



narrow strip of moderately elevated 
Pulangui, flowing northwest in a 
valley with long, gentle slopes up 
to Mount Katunlud on the west, 
but with very rugged hills on the 
east. In the upper part of its 
course its tributaries join it, mak- 
ing an acute angle upstream in 
the normal way, but below the 
confluence of the Manguina and 
the Tagoloan Kivers the latter 
enters a narrow gorge with only 
one side stream coming in from 
the west in a distance of 19 kilo- 
meters, but with 24 entering from 
the east and all of these approxi- 
mately at right angles. ( See Plate 
III.) This remarkable drainage 
arrangement must be due to fault- 
ing and joining. There undoubt- 
edly was a local uplift east and 
west across the courses of theee 
rivers, after they had become well 
established. The country to the 
south of their headwaters be- 
comes more open and flat. 

(4) The Tumaga River is very 
small and scarcely navigable above 
1^ kilometers from its mouth. It 
is mentioned here because of 
its importance to Zamboanga as 
a source of potable water. It is 
not much over 32 kilometers in 
length and very narrow and 
shallow. It rises on the slopes of 
Mount Panubigan, flows about due 
south over schists and along their 
strike until it issues from the 
Zamboanga gorge onto the plain, 
where after a short distance it 
turns sharply to the east. Just 
what has so sharply deflected this 
stream is not apparent, unless, 
perchance, during a freshet it de- 
posited on the plain near Zam- 



country from the headwaters of the 




352 



SMITH. 



/•i^' 



boanga a great load of detritus which afterwards turned its course to the 
east. A slight warping of the, coastal plain 
would also explain this eccentricity. 

THE INTERMEDIATE UPLANDS. 

Under this heading is comprised all the ter- 
ritory lying between the plains and the Cordil- 
leras. With the exception of the range east of 
the Agusan Eiver, the Kulingtang Mountains 
south and east of Lake Lanao, and the range 
of low mountains in the Zamboanga Peninsula, 
there is no distinct cordillera in Mindanao and 
the ranges cited are not of sufficient height and 
continuity to place them exactly in that category. 
By far the largest portion of the area of 
Mindanao then should be considered as simply 
i upland country. We have seen that the central 
I plain of Mindanao is fairly low. The prin- 
I cipal upland areas, therefore, are the peninsular 
g portion south of Lakes Liguasan and Buluan, 
o known in part as the Tiruray Tableland; the 
w Lake Lanao upland follows; then, third, prac- 
^ tically all of Zamboanga Peninsula ; and, fourth, 
§ the tract in which rise the Sahug, Sabul, and 
a Ae^usan Rivers. 

i 

^ THE TIRURAY TABLELAND. 

o The Tiruray tableland is practically un- 

V explored. I have been on the edge of it near 
^* Tamontaca and have sketched the southern sky- 
^ line to the south. (See fig. 1.) Several prom- 
inent points, Mounts Blik, Itim-Itim, etc., rise 
above the tableland; their elevation is between 
366 and 457 meters. The people inhabiting 
this territory are Tirurays, a degenerate band 
of Moros, Manobos, and Tagabilis, all pagans. 

THE LAKE LANAO UPLAND. 

This tract, bounded on the north by Iligan, 
on the south by lUana, on the west by Panguil 
Bay, and on the east by the Kulingtang Eange, 
is one of the most interesting in all the Philip- 
pines. The average elevation is about 610 
meters. Lake Lanao is 686 meters above the 
sea, but some of the country to the west is lower, 



qq\V 



.- ii: K 



MINDANAO AND SULU: II. PHYSIOGRAPHY. 



353 



while certain of the most prominent points, such as Mount Gurayu, are 
over 1,220 meters. It is for the most. part treeless, and covered. with a 
rich, red soil. Cogon grass covers the greater part of the slopes. 

Physiographically, the Lanao upland is unique. It bears - a very 
striking likeness to the elevated prairies of the middle west of the 
United States. A profile from Camp Overton to the lake and south 
to Malabang is shown in figure 2. One very striking feature of this 
region is Keithley escarpment, a bold ridge with an abrupt face to the 
north and a long, gentle slope southward to the lake. A good idea 
of the topography of this section is obtained from the panorama. (See 
Plate IV.) It is the belief of many people who have seen this country 
that Lake Lanao occupies a crater. The topography (see fig. 3) of 
the country in general and particularly that of this escarpment does 
not, to me at least, lend much encouragement to this supposition. It 
is my belief that it is simply a basin between two well-defined mountain 
ranges, which has been dammed by lava flows and an unusual accumul- 
ation of wash from the hills. 




FiQ. 3. — Sketch of Topography near Lake Lanao, Showing the Keithmiy 

ESCARPMENT. 



364 SMITH. 

To the west of the lake there are long, grassy slopes but the country 
is sparsely inhabited. However, on the east there is a considerable tract 
of perfectly level, somewhat swampy land, with a fairly bold, heavily 
wooded range behind it. A large population of Moros is found on this 
side of the lake. These are known as the Lanao, or lake Moros. This 
region may be called the last great stronghold of the Moros. I have 
already given some account of them in Part I of this work.^ 

This upland region is the most favorable in all Mindanao for white 
settlements; it is high and cool, and possesses a wonderfully rich soil, 
particularly adapted to coffee growing. The country is well suited to 
stock raising. 

THE ZAMBOANGA PENINSULA. 

The Zamboanga Peninsula owes its origin to a sharp upward flexing 
of^ sedimentaries. A veneer of volcanic material overlies them. It is 
strictly a cordillera, but only of moderate elevation. It is practically all 
upland country, but well dissected. The chief characteristics of this 
region can be enumerated as follows: (1) It is long and narrow; (2) it 
has a moderate elevation; (3) there are a great number of indentations 
in the coast line; (4) it has a small development of coastal plain; (5) it 
is topographically youthful, with occasional longitudinal streams and 
a great number of short, swift, consequent streams at right angles to 
the main trend of the peninsula; and (6) superimposed high points 
rise above the general uplands, such as Mounts Panubigan, Dapiok, and 
Malindang. 

The region is inhabited almost exclusively by Subanuns, a non- Chris- 
tian hill people who have evidently fallen back before the Moros. They 
live in caingins (small clearings), where they cultivate camotes as their 
principal crop. They are very primitive, but in my opinion possessed of 
many good qualities. They are not, where I have encountered them, at 
all warlike.^ 

THE DAVAO UPLAND. 

This upland area is not very large. It comprises the country between 
the Apo Eange and the Agusan Eiver. Its most prominent feature is 
a fairly well-defined ridge which trends east and west and northeast. Its 
two highest points are Mounts Panombayan and Kuanabayan. Very little 
is known about this country. The only scientific observers who have 
traveled through it are Montano, who paid more attention to the character 
of the people than to the geology, and Ickis and Goodman, formerly of 
this Bureau. Their observations have already been published in Part I 

*Th%8 Journal, Sec. A (1908), 3, 473. 

•Christy, E. B. The Subanuns of Sindangan Bay. Bur. Sci. Sub., Div. of 
Ethnol. (1909), 6, pt. 1. 



MINDANAO AND SULU : II. PHYSIOGRAPHY. 



355 



of this report. Montano's remarks relate only to the character of the 
rocks and hence should properly be left to the next chapter. 

From Goodman's description it is^evident that the country is quite 
rough and heavily timbered, with occasional clearings in which the long- 
haired Mandayas live. 

THE CORDILLERA. 

As I have said, the only cordillera proper is that east of Agusan River. 
Very little accurate information about this region is at hand. Ickis 
crossed it from Talacogon to Lianga, but his untimely death prevented 
his making a report on what he saw. There are two other prominent 
trails across this cordillera, one from Compostela on the upper waters of 
the Agusan to Cateel, the second the route taken by Goodman from San 
Juan to Mati. This region is inhabited largely by Mandayas. 

Exceedingly rugged topography would be expected in this region, be- 
cause it is on the edge of the great continental horst. It is one of the 
most humid districts of the entire archipelago. The rainfall at Caraga 
on the eastern side of the cordillera from September, 1902, to August, 




Fig. 4. — Outline Map of Mindanao Showing Cobdillbras and Principal Drain- 
age Lines. 



356 , SMITH, 

1903, was 3,165 millimeters. This was exceeded in only two other places 
in the Philippines, namely, at Borongan, east coast of Samar, and Ma- 
sinloc, west coast of Luzon. 

From the fragmentary notes we have on this portion of the country I 
infer that the rocks of this cordillera are much the same as those in the 
Cordillera of Luzon. I do not expect to find the rocks of Mindanao any 
older than those of other parts of the Islands, but possibly somewhat 
younger. 

Figure 4 gives a provisional idea of the main tectonic lines, to which 
is added the hydrography as we now know it. 

LAKES. 

There are in all thirteen lakes known in Mindanao. In order of size 
they are: Lanao, Liguasan, Buluan, Linao, Mainit, Kadagan, Malanao, 
Leonard Wood, Balut, Dapao, Butig, and Munay. 

Lake Lanao. — ^This has already been referred to under another head. 

Lakes Ligvxisan and Buluan are simply the remnants of a greater 
body of water which formerly occupied all the central low country. 
Lieutenant Van Horn, United States Army, made an exploring trip 
in 1902 from Cotabato across the Tiruray table-land to Makar and 
back by way of the valley between Mounts Malibatu and Matutan. He 
was of the opinion that the Eio Grande once flowed through Lakes 
Liguasan and Buluan along the depression now followed by the trail 
through Talik and Tambatu. This, I think, is very probably correct. 
These lakes are very shallow, being little more than swamps. Lieutenant 
Van Horn in his manuscript report to the Adjutant-General says:^^ 

Our route was through the Buluan River connecting the two lakes. The 
entrance to the river from lake Buluan was not more than a meter wide on 
account of being choked up by floating islands of grass, but it widened out to 
about forty meters for about 6 kilometers. It was hard to distinguish the river 
from the numerous esteros. After 7 kilometers we came to some solid ground 
on the left bank where there was a tiangi or market. After this we could see 
no solid ground on either bank until I reached the Rio Grande. Buluan river 
is very swift, so swift in the channel we came down that vintas must go up by 
another route. In one place we passed over a small fall about three-quarters of 
a meter in height, the turn and the bottom being very sharp and we going so 
fast that our vinta was forced through the high grass on the bank about 9 meters. 
Entered Lake Liguasan about 1 P. M. The lake was very much the same as 
the surrounding country, a few patches of open water, large floating islands, 
and the rest covered with lily pads and water cabbages. A small channel 
was made by boats cutting their way through. The character of the lake changes 
with every wind. 

Lakes Linao and Kadagan, — These two lakes are not much more 
than marshy areas where the Agusan Eiver has spread out over a large 
tract of low ground. There are three lakes marked on the latest map 

^•Van Horn, Lieut. R. C. MS. Report to Adjutant-General, United States 
Army, Philippines Division, January, 1902. 



MINDANAO AND SUJiU: JI. PHYSIOGRAPHY. 357 

issued/^ called Dinagat, Dagun, and Linao. The first two are really 
one and the other is only a few kilometers farther up the river. These 
bodies of water are very shallow in places, in others there are trees 
partly submerged. They are navigable for hancas (dugouts)^ a^d in 
one place a channel, navigable for light (3 feet) draught launches, has 
been located, beginning near Bunauan and ending in the main river. 

Lake Mainit, — ^This is a smaller lake, more pear-shaped than it appears 
on the older maps, about 9.5 kilometers in diameter, situated near the 
northern point of Surigao Peninsula. Its name, meaning hot, its shape, 
and the more or less high land around it point to the possibility of its 
being a caldera, 

Mr. Montano,^^ in 1881, wrote as follows concerning Lake Mainit: 

The large lake of Mainit, situated at the center of the peninsula, at an altitude 
of 40 meters, seems to be the crater of an ancient volcano; it is circular, very 
deep, and its banks are very steep; it is surrounded by high mountains, where 
hot springs abound. He calls attention in another paragraph to some limestone 
caves on the east side of the lake. 

Besides the moderately large lakes already discussed, there are several 
smaller ones, like Malanao and Balut near Cotabato; these are little more 
than ponds in the chance depressions in the topography. Lakes Malanao 
and Balut owe their existence to small synclinals in the sandstones and 
shales just north of the Kio Grande. The same is true of Butig Lake 
near the southeast corner of Lanao. 

Lakes Munay, Dapao, and Nonungan are due to depressions in the 
basalt flow which covers most of that region. 

Information concerning Lake Leonard Wood is confined to a brief 
mention in a report to the Adjutant-General, United States Army, by 
Capt. C. C. Smith, Fourteenth Cavalry, concerning an expedition made 
in June, 1904, from Misamis to Dumaquilis Bay. This lake has an 
altitude of 889 meters, is shaped like a figure 8, and roughly is 8 
kilometers long by 3 kilometers wide. 

These bodies of water have played a most important part in the history 
of Mindanao. The Moro is by nature warlike. He has lived largely near 
the sea, alternating between the peaceful occupation of fishing and the 
more exciting sport of plundering villages and taking of slaves. With 
the coming of the Spaniards, Visayans, and Chinese, and their settling 
along the coast of Mindanao, numbers of the Moros retreated into the 
interior as the Subanuns and other primitive peoples had done before 
them. Naturally, they settled around the inland bodies of water, Lanao 
and Liguasan, where they could resume their old manner of living with 
as little interruption and as little change as possible. Therefore, today 
the last great stronghold of the Moros lies surrounding the upland lake of 

"World Book Co. (1907). 
, ^Loc, cit., p. 288. 



358 SMITH. 

Lanao. Not that Moros are not to be found in force elsewhere, but here 
and in Basilan are the only places where they still evade complete 
pacification. 

The Moros around Liguasan are more peaceful because they are more 
easily reached. They are not found east of the Apo-Matutan Kange. 

It must not be inferred from what has just been said that the Moros 
pay no attention to agricultural pursuits. On the contrary, they devote a 
great deal of time to the raising of rice in the Cotabato Valley and on 
the fringe of low land on the east side of Lake Lanao. In Sulu the 
Morons principal farm product is tapioca. 

VOLCANOES. 

There are no active volcanoes at the present time in Mindanao. One, 
Mount Apo, has some vents on the eastern slope which emit steam and 
sulphurous fumes, but the mountain is in no sense active. Vulcanism, 
which is now merely an incident in the life of the island, was at one 
time the dominant feature. This period of volcanic activity was prob- 
ably the Pleistocene, when Mindanao was taking on its final shape; then 
Mounts Matutan and Apo near Davao Gulf, the Buldun Mountains south 
of Lake Lanao, Mount Malindang west of Panguil Bay, and the moun- 
tains near Lake Mainit were in all probability belching forth lava and 
ashes. Now these are all quiet ; if not extinct, at least dormant. 

If we look closely at the map of Mindanao, we find that the volcanic 
centers which seem scattered indiscriminately over the country in reality 
lie along certain definite lines which intersect and form a triangle. These 
trend N. 6° W., N. 68° E., and N. 53° W. Catarman, the active volcano 
on the Island of Camiguin, is located at the northern end along the first 
line and Mount Apo at the southern. Mounts Panubigan, Tres Eeyes, 
Sugarloaf, Malindang, and Camiguin are situated on the second line 
with Lake Mainit (caldera) and its hot springs almost on it. The third 
line runs through the Lanao cluster of old cones and southeast to Mount 
Apo. Mount Apo and the Kulingtang cones preserve much of their 
former shape, the others are much more worn and dissected by erosion, 
so that they are now little more than volcanic stocks. Mount Malindang 
is a very imposing pile of andesite and basalt, rising to the commanding 
elevation of 2,800 meters from the shores of Iligan Bay. It. was last 
ascended in 1906 by Lieutenant colonel E. A. Mearns, United States 
Army, retired, and his party. 

THE KULINGTANG RANGE. 

As the traveler rides toward the south down the long grassy slope from 
Camp Vicars on the edge of Lake Lanao he has before him a magnificent 
panorama of this long range with its numerous burnt-out craters arranged 
in step-like fashion like so many blown-out blast furnaces. My photo- 
graph of this range unfortunately was spoiled. 



MINDANAO AND SULU: II. PHYSIOGRAPHY. 359 

THE APO VOLCANIC CLUSTER. 

Mount Apo is the highest peak in a volcanic cluster on the west side 
of Davao Gulf. This peak, although higher, is not nearly as large as 
the old crater to the northeast. As I have said elsewhere, I seriously 
doubt if the present Mount Apo ever has been an active volcano. The 
old mountain, the two highest points of which are called Culelan and 
Pumantigan, was doubtless the center of a very great disturbance in 
former times, as the nature of eruptive material indicates an explosion 
or series of explosions which must have devastated the country for many 
miles. The photograph (frontispiece) was taken from Digos. 

I have no first-hand information concerning Catarman, the only active 
volcano which might be considered as belonging to Mindanao, it being 
on the little Island of Camiguin off the Misamis coast. An extensive 
account of this volcano and the eruption of 1871 has been written by 
Father Maso, S. J.^^ 

SHORE FEATURES. 

Coral reefs, — There are a number of phenomena along the shore 
which vitally affect the life and customs of the inhabitants of the island. 
Coral reefs are perhaps the most conspicuous of these features. The 
latter are found in scattered patches all around the island. However, 
the growth is not great on exposed coasts like the stormy eastern one, 
but in sheltered bays like Sibuguey and Dumankilis it is exceedingly 
difficult to navigate because of the reefs. They act as danger spots to 
the navigator and as a protection to the people on shore, for besides 
making it difficult for invaders to land they break the seas in the time of 
great storms. A remote benefit due to reefs, remote because of the 
exceedingly long period of time which must elapse before it can be 
realized, is that they aid in the growth of the land. The coral polyps 
grow up to a limiting line which is the level of the sea, and upon this 
platform is deposited the detritus borne down from the mountains. In 
time this will be dr}' land. If the whole body of the lajid is rising, as 
is the case with much of Mindanao at the present time, this growth 
proceeds with comparative rapidity. The prominent part that reefs 
have played in the formation of Mindanao can readily be realized when the 
raised reefs in many parts, particularly those found following the valley 
of the Eio Grande, are seen. 

Terraces. — One of the most striking of all physiographic phenomena, 
the significance of which is not always appreciated by the layman, is 
the terracing along the seashore and on the sides of the valleys. 

"Mas6, Rev. M. Saderra, S. J. Volcanoes and Seismic Centers, 208-216. Phil- 
ippine Census. 



360 SMITH. 

Terraces originate in several ways and according to their origin we 
discriminate the following : 



1, Flood-plain terraces. 

2. Those due to inequalities of hardness of strata. 

1. Wave built. 

2. Wave cut. 

Lake terraces The same as marine. Usually on a smaller scale. 



River terraces- 
Marine terraces.. 



I have already alluded to the terraces of the Cotabato Yalley. These 
are in part marine wave cut, in part river cut, and also due to the eleva- 
tion of reefs. However, the most striking terraces are those along the 
present seashore. Exceptional examples are to be seen. At Point 
Blanca, on the northwest coast of Mindanao, there is a raised delta of 
an old river. It is sliced off at the sea margin so that the structure is 
clearly revealed. A sea cliff exists in this soft material of from 6 to 9 
meters elevation (estimated from the boat). Formerly, the lower end of 
this delta was at the level of the sea. The structure as now apparent is 
that of a typical delta with the cross-bedding always found in such forma- 
tions. 

The second striking example is to be seen in Makajalar Bay. Here 
the terracing is very pronounced. I am indebted to Miss Eleanor P. 
Bliss for the following sketch made on a visit to this locality in 1907. 
(See ^g'. 5.) 

The meaning of these terraces, raised deltas, and beaches is that 
Mindanao, in some quarters at least, is, rising. Very accurate coast and 
geodetic work, such as is being carried on now, and a comparison of the 
present with a resuryey fifty or a-hundred years hence, might reveal a 
measurable increase in area. 

SULU. 

In the Sulu Archipelago vulcanism has played a most prominent 
part and for the most part is of more recent date than that of Mindanao. 

On the Island of Basilan there are many extinct cones and more on 
Jolo, iaccompanied by many hot springs. On Tawi Tawi the signs of 
recent vulcanism are not so pronounced, and as far as I know this island 
is largely made up of sedimentaries. Siasi is practically a partially 
wom-dowii volcanic stock. In all these islands we have no distinct 
Cordilleras, but numbers of more or less isolated cones which do not 
appear even to have linear arrangement, although a more detailed survey 
might reveal some such system. Subsequent erosion has produced in 
these islands a topography which is peculiar and very pleasing in its 
long, gentle slopes and beautiful curves. Jolo is only partially wooded, 
there being long stretches of fields covered with waving cogon with here 
and there cleared patches on which the industrious Moro has cultivated 
the tapioca plant with its striking pink blossom. The soil is deep, rich, 
red, and exceptionally fertile, and would be excellent for sugar. 



MINDANAO AND SULU: II. PHYSIOGRAPHY. 



361 



On the 



Island of Job tapioca and coconuts are the chief products. 
Basilan has several rubber plantations. 
Siasi and Lapac, just across a narrow 
strait from it^ are now being given 
over to sago, coconuts, and tapioca. 

The striking feature about this little 
archipelago is of course the multitude 
of inlets with innumerable little bays 
and proportionally great length of 
coast line. These features make this 
one of the best possible resorts foi* 
pirates and hardly a year passes but 
the United States troops have some 
slight encounters with them. How- 
ever, piracy has now practically ceased. 

There are three physiographic phe- 
nomena which I desire to call attention 
to before leaving this subject. They 
are: 1. Atolls; 2. Monadnocks; 3. 
Crater lakes. 

Atolls. — I shall not repeat here n 
description of an atoll. It is sufficient 
to say that there are several more or 
less perfect ones in this region, some 
with lagoons of considerable size in the 
center^ others with only shallow de- 
pressions, probably a result of eleva- 
tion and filling of the former basins. 

Monadnochs. — I have used this 
word for want of a more concise term 
to describe the phenomena seen on 
some of the islands, particularly Ma- 
rongas, a very small island just north 
of the town of Jolo. A photograph 
of these peculiar humps sticking out 
of the sea is shown on Plate V. They 
are simply due to hard, resistant basalt 
dikes which have intruded the sedi- 
ments and have withstood sea action. 

Crater lakes. — There are several of 
these on the Island of Jolo. (Plate 
VI.) Silt Lake near the constricted 
portion of the island is particularly 
noteworthy. However, by far the 
most striking of these lakes occur on 




362 SMITH. 

Cagayan Sulu, where on the southern side there are three almost circular 
in form. These are so near the seashore that the sea has broken through 
the rim of one. 

This closes my brief discussion of the physiographic units in the 
Island of Mindanao. It can be seen, then, that this is one of the newest 
and most interesting portions of the whole Philippine Archipelago, a 
region of slow but continued change. Possibly in time there will be an 
unbroken land bridge between Borneo and Mindanao. At the conclusion 
of the next paper, when I shall have discussed the geology, I shall take up 
the broad questions of human response to natural conditions and attempt 
to point out some lines along which development is likely to proceed in 
this great island. 



ILLUSTRATIONS. 



Plate I. Mount Apo as seen from Davao Gulf. 
II. A meander of the Rio Grande. 

III. Topography along the Tagoloan River. 

IV. Panorama of the Lanao upland. 

V. Islands off the north coast of Jolo. 
VI. Map of the Island of Jolo. 

TEXT FIGURES. 

Fig. 1. Sky line south of Cotabato. 

2. Profile from Camp Overton to Malabang. 

2. Sketch of topography near Lake Lanao, showing the Keithley escarpment. 

4. Outline map of Mindanao, showing cordilleras and principal drainage lines. 

5. Terraces at the mouth of the Cagayan River. 

99514 5 363 






^ ...' '.- 



o 







9 9 a^a *»° •!! 






> 


^a 


i4 




M 
O 

U2 




§ 

t-5 


■ ■.'■''■i 


(1. 









0* 



Smith : Reconnoissance of Mindanao and Sulu.] 



[Phil. Journ. Sci., Vol. V» No. 5. 




PL-AXE III 



Smith : Rkconnoissance of Mindanao and Sulu.] 



[Phil. Journ. Sci., Vol. V, No. 5. 




p».i_A-rE: IV. 






^ 




REVIEWS. 



Allen's Commercial Organic Analysis, Volnme I. Edited by Henry Leffmann 
and W. A. Davis. 4th ed. Cloth. Pp. x+576, 86 illustrations. Price $6 
net. Philadelphia: P. Blakiston's Son & Co. 1909. 

This admirable volume is composed of the following divisions: (1) 
Introdiu3tion by William A. Davis, treating with general methods, includ- 
ing preliminary examination, specific gravity, changes in physical state, 
optical properties, spectrometers and spectrographs, polarimeters, ar- 
rangements for maintaining known constant temperature, analysis, 
moisture, crude fibre and ash, and action of solvents, 83 pages; (2) 
Alcohols, by C. C. Jones, 47 pages; (3) Malt and malt liquors, by 
Julian L. Baker, 31 pages; (4) Wines and potable spirits, by C. C. Jones, 
40 pages; (5) Yeast, by Emil Schlichting, 21 pages; (6) Neutral alcoho- 
lic derivatives, by Henry Leffmann, 54 pages; (7) Sugars, by E. Frank- 
land Armstrong, 119 pages; (8) Starch and isomers, by E. Frankland 
Armstrong, 59 pages; (9) Paper and paper-making materials, by E. W. 
Sindall, 19 pages; (10) Acid derivatives of alcohols, by Henry Leffmann, 
82 pages ; an Appendix, 1 page, and an Index, 6 pages. 

The book is a useful compilation for analysts and in all but a vei*y few 
instances is well up to date. It is but natural for each individual 
worker to find something to criticize and to discover omissions which 
he would like to have included in the work. On page 258, under the 
Eimini reaction for formaldehyde, is omitted the very satisfactory test 
given on page 185 of Bulletin No. 107, Bureau of Chemistry, United 
States Department of Agriculture. 

Under the heading "Detection and estimation of chloral,^^ the delicate 
test of Covelli, depending upon the coloration of fatty oils (Chemiker 
Zeitung (1907), 31, 342), is omitted. It is possible that these methods 
appeared after the book had gone to press (the first appeared in 1908 and 
the second in 1907), although this does not seem to be the case, for the 
section upon potable spirits is completely up. to date and contains refer- 
ences to most improved methods of analysis (p. 190) and their inter- 
pretation (p. 202). The considerations of mannitol and the methods 
for its estimations and detection does not appear in this volume while in 
the former editions several paragraphs are , given on this subject. 

The typography and binding leave little to be desired, although a few 
errors are observed. This edition should be in every chemist's library. 

H. D. G. 



366 REVIEWS. 

Allen's Commercial Organic Analysis, Volume II. Edited by Henry Leffmann 
and W. A. Davis. 4th ed. Cloth. Pp. x+520. Price, $5 net. Philadelphia: 
P. Blakiston's Son & Co. 1909. • 

The previous editions of this work have been so eminently serviceable 
that analysts will naturally direct their attention to the fourth and latest 
edition. Much of the subject-matter of the present volume appeared in 
Volume II, Pai*t I, of the third edition of ^^Commerciar Organic 
Analysis/^ but the analytical chemistry of explosives^, which was partly 
included in that volume^ has been omitted. 

As is to be expected in a book of this kind, written by several contrib- 
utors, different parts are of very different merit. In the opinion of the 
reviewer the chapters on Special Characters and Methods, by Leonard 
Archbutt, and Linseed Oil, by C. A. Klein, are particularly valuable- 
The analyst will welcome the abundance of new material to be found in 
Mr. Archbutt^s chapter. The new data on tung, soja-bean, coconut, and 
candlenut oils will be particularly welcome to analysts in the Orient. 
The recent work by Mr. H. S. Walker of this laboratory on coconut oil 
is included. 

The chapter on General Properties and Analytical Methods by C. 
Ainsworth Mitchell does not appear to have been brought up to date as 
thoroughly as the other chapters. The author fails to describe the Hanus 
modification of the Wijs method for iodine absorption, although iodine 
numbers obtained by this method are quoted in otlier chapters of the book. 
Inasmuch as this method is widely used and is recommended by Leach, 
Wiley, and other authorities, the reason for its omission is not clear. 

Altogether the book is a splendid compilation and an indispensable 
asset for a well-equipped laboratory. 

B. T. B. 



THE PHILIPPINE 

Journal of Science 

A. Chemical and Geological Sciences 
AND THE Industries 



Vol. V DECEMBEE, 1910 No. 6 



PHYSICAL AND CHEMICAL PROPERTIES OF 
PORTLAND CEMENT. —PARTS I AND II. 



By W. C. Keibling and F. D. Reyes. 
{From the Chemical Laboratory, Bureau of Science, Manila, P. I.) 



INTRODUCTION. 



In a paper on the subject of Portland Cement which appeared in a 
previous number of this Journal ^ the efficiency of modem cement spec- 
ifications and standard methods of testing were discussed, and the effect 
of climatic and atmospheric influences on certain brands of commercial 
Portland cement noted. Incidentally, exception was taken to the pub- 
lished statements of several authorities regarding the significance and 
value of the specific gravity and soundness tests. It was also pointed 
out that "seasoning" could injure as well as benefit cements and almost 
incredible changes in the physical properties of some cements produced 
by slight changes in temperature, or by aeration, were recorded. In con- 
clusion, the authors suggested the characteristics which cements should 
possess in order to give the best efficiency in tropical countries, and stated 
that investigations would be continued on samples of many grades of 
Portland cement in order more thoroughly to test the soundness of their 
deductions. 

Our main efforts in continuing this work have been directed toward a 
study of those characteristics of Portland cement regarding which there 

1 Reibling, W. C, and Salinger, L. A., This Journal, Sec. A (1908), 3, 137-185. 
99667 367 



368 REIBLING AND REYES. 

exists the greatest amount of misconception and diversity of opinion, our 
object being to assist in the universal effort to improve cement specifica- 
tions. It is our desire to affiliate our efforts with those of the American 
Society for Testing Materials and with similar associations which are 
endeavoring to formulate cement specifications so drawn as to guarantee 
the manufacture and use of Portland cement of the quality sought for. 

The quality which we need in cement is constancy; constancy in satis- 
factory setting properties^ in volume, in strength, and in sand-carrying 
capacity. The demands of modern construction work have also made it 
^^of vital importance that this material should not only harden rapidly 
and attain a great strength, but what is far more essential, that it maintain 
this strength." ^ However, sixty years of strenuous effort have failed to 
perfect methods of testing or to produce specifications which will insure 
a cement of this quality, or correctly to classify the various products in 
their true order of merit. 

Our investigations were conducted on many grades and brands of mate- 
rial. Few places in the world offer a greater variety of cements than 
Manila. Portland cement is not manufactured in the Philippines and 
our tariff and location make it possible for cement manufacturers through- 
out Europe, America, and the Orient to place their products on sale here 
at almost identical prices, so that we had at our disposal brands of Port- 
land cement manufactured in all parts of the world and representing 
almost every class of raw material, process of manufacture, and degree 
of sintering. These various products were sent to the laboratory, tested, 
their characteristics noted and larger representative quantities then ob- 
tained for research purposes. 

Each brand was mixed thoroughly and then preserved in 10-liter, 
air-tight bottles until used. It was analyzed chemically and the physical 
tests made which are outline in the specifications of the United States 
Army (1902), the American Society for Testing Materials (1908), and 
the Philippine, Government (1908),^ the strength and soundness deter- 
minations being continued for one year. Each brand of cement was then 
aerated, and from time to time repeatedly subjected to a like physical 
examination. Tests of aerated samples were accompanied by similar 
tests on the unaerated cement. These constituted only the foundation 
for a more thorough investigation of the cause and significance of all 
results not readily explainable. Microscopic, chemical, and calorimetric 

2 Humphrey, Richard L., Am. Soc. Testing Mat. (1902), 2, 160. 

' In all cases the temperature during manipulation and exposure was main- 
tained between 25 and 30 degrees. The Philippine Government specifications of 
1908 were identical in all other respects with those of the American Society, 
except that a natural Philippine (Tarlac) sand, screened through a 20- and on a 
30-mesh sieve was used instead of Ottawa sand. Recently standard Ottawa sand 
has once more been adopted. 



PROPERTIES OF PORTLAND CEMENT. 369 

tests ; compression determinations on concrete cubes, mortar cylinders and 
the broken tensile-strength briquettes ; expansion bars ; air, moist-air, and 
sea-water exposures; and acid and alkali treatments, were all utilized. 
After working on the commercial product, we continued the investigation 
on the nonaerated clinker received from manufacturers in Europe and 
China. Finally, one of the authors visited a few cement mills, where 
every courtesy and assistance were given to him and where he was enabled 
to secure valuable information and collect special material for this work. 
For the sake of brevity and clearness the work which we have done 
will be priesented, as far as possible, as an abstract of the results. 

CHEMICAL ANALYSES. 

Table A (Appendix I) gives the chemical analyses and physical 
characteristics of typical samples of various brands, the requirements and 
methods of the American Society for Testing Materials being used. 

The chemical analyses of the same brand of cement taken from time to 
time during the past five years usually have shown very little variation, 
and the analytical work of this Bureau has been found to agree very closely 
with similar records taken from factory reports on identical material. 
Despite this uniformity in chemical composition, we have at times encoun- 
tered much variability in physical properties. Even in the same shipment 
we have met with sound and unsound, and slow and quick-setting cements, 
as well as cements which developed both low and high seven-day tests, 
accompanied by good increases, or slight decreases in later determinations 
of strength. 

On the other hand, some brands of Portland cement have shown the 
most remarkable uniformity in physical properties. The significance of 
chemical composition and the relationship between chemical and physical 
properties are given under the specific heads of the discussion. 

Part I. 

FREE LIME IN PORTLAND CEMENT. 

As our investigations progressed it became evident that the physical 
properties of Portland cements were influenced to a marked degree not 
only by the amount of free lime, but also by its condition, that is whether 
this calcium was present as the hydroxide, oxide, or as the latter heated 
to a degree of incipient fusion. 

The most diverse opinions have been expressed concerning the ques- 
tion of the existence and importance of free lime in Portland cement. 

Alfred H. White* says: "It is rather generally, although by no means univer- 
sally, assumed that the lime must be in a state of combination, but the evidence 
for this belief is almost entirely indirect." 

*Journ. Ind. d Eng. Chem. (1909), 1, 6. 



370 REIBLING AND REYES. 

D. B. Butler* makes the following statement: "Unstable lime compounds, or 
rather lime compounds which, after a more or less prolonged lapse of time, 
slake with considerable expansion, there certainly are, and these are doubtless 
answerable for the unsound cements which are sometimes met with, but anyone 
with experience in such matters will appreciate the fact that free lime, i. e., 
calcium oxide, can not exist in contact with argillaceous or acid compounds at 
ordinary cement kiln temperature." 

W. Michaelis, Sr.,® writes as follows: "Even by the clinkering process the 
inside of the coarser grains of calcined lime can never be reached by the fusing 
silicates surrounding them. The powdered clinker therefore must contain a 
certain amount of uncombined lime." 

The more direct evidence of chemical, thermochemical and microscopical ex- 
periments has also led to diverse conclusions. 

W. Richter^ believes that all lime in cement is combined. Steuer,^ Wormser 
and Spanjer,® and Hart,^® on the contrary, state that they have extracted 44, 
26.6, and 30.34 per cent, respectively, of calcium oxide free from combination, 
from Portland cement. 

There is also a diversity of opinion in regard to the physical and chemical 
properties of free lime and its effect upon the properties of the cement. 

Hart ^ states that the reason why 30 or 34 per cent of free lime can exist 
in Portland cement without injury to the product is because this lime is vitrified 
and crystalline, differing physically from amorphous lime. He believes that on 
hydrating, this vitrified lime hardens under water and considers the balance of 
the cement material as ballast, as slag which does not harden. Others ^ regard the 
lack of much free lime in the rotary clinker as detrimental. They believe that the 
old set-kiln process was best, even though it produced a cement which contained 
much free lime. They state that under the modern practice of ageing, grinding, 
and mixing, a considerable amount of free lime is taken care of and that, although 
the rotary kiln has reduced the percentage of free lime, it has worked an injury 
because the increase in the silica and alumina ratio has resulted in a cement 
that fails to maintain its high early strength. 

E. D. Campbell,^* assumes the opposite to be true and writes as follows: "Free 
lime in Portland cement will not only not be slaked during the mixing and setting 
of the cement, but will not become completely hydrated even when the cement 
is immersed in water until about fourteen day» have elapsed. The result of this 
gradual slaking is to produce abnormal expansion of the cement. The expansion 
due to free lime slaking in the air may become so great after several months as 
to cause complete disintegration.*' 

E. B. McCready's ^* views are similar. He states that : 

"When the amount of uncombined lime is not enough to cause disruption, 
the strain exerted simply weakens the specimen so that it is more readily broken 

'^ Official Correspondence. 
"" Cement d Eng. News (1909), 21, 344. 
"^ Thonindustrie Ztg. (1883), 46. 
»/&wi. (1899), 604. 
^Ihid, (1885). 

^o/6td., 659, 770, 853; (1900), 188. 

" Loc. cit. See also Blerninger, Albert Victor, The Manufacture of Hydraulic 
Cements, (ld04), 215; Bull. Oeol. Surv., Ohio (1904) IV, No. 3. 
^Journ. Amer. Chem. Soc. Ahs. (1908), 2, 2983. 
^Journ. Amer. Chem. 8oc. (1906), 6, 1298. 
^^Amer. Soc. Test. Mat. (1907), 7, 375. 



PROPERTIES OF PORTLAND CEMENT. 371 

in the testing machine. The rate at which this disruptive strain increases in 
different samples under various conditions of burning, grinding, and testing is 
the kernel of the nut which we ought to crack before placing too much reliance on 
values deduced simply from analyses or percentage of gain/* 

Heretofore, it has been impossible to make other than a very limited study 
of free lime in Portland cement. The few reliable methods*' for its detection 
and estimation were so difficult and tedious as to be impracticable for general 
application. They also failed to reveal the physical characteristics of lime so 
detected. 

However, Alfred H. White ^^ has recently described a quick, delicate, and 
reliable method for the detection of free lime, the development of which promises 
to be valuable. Unfortunately, this method has not yet been made quantitative, 
but its application enables us to determine the relative amounts of calcium oxide 
present and to distinguish between that which is sintered, nonsintered, or hydrated. 
White has not had time nor opportunity to develop all of the possibilities of his 
discovery; and those deductions which he has made from the application of this 
test on commercial Portland cements have proved faulty in some cases because of 
his assumption of a somewhat universal misconception, that ^'cements which pass 
a perfect boiling test may safely be assumed to contain no free lime"" and because 
the phenomenon of the formation and development of his so-called calcium phe- 
nolate crystals was not sufficiently understood. 

This laboratory has employed White's test in the study of the cements 
which have come to hand and as a consequence it has acquired a funda- 
mental importance in our interpretation of the causes of many otherwise 
inexplicable phenomena. The proofs of the presence and the effect of 
free and hydrated lime in commercial cements are so conclusive that a 
somewhat extensive account of our investigations in this respect is 
necessary. 

CALCIUM HYDROXIDE-PHENOL. 

Correctly to interpret the significance of the crystals described by 
White and termed by him "crystalline calcium phenolate/^ we investigated 
their formation^ habit^ and composition. 

Phenol will not react with calcium oxide or with calcium hydroxide 
except in the presence of water. Calcium hydroxide will combine with 
phenol and water to form characteristic, long, slender crystals which are 
extinguished by parallel and visible with crossed nicols. In an excess 
of phenol, the length to which these crystals may grow depends upon 
the amount of lime and water available at all stages of the reaction. By 
carefully controlling these factors we have succeeded in obtaining crystals 
2.0 millimeters long. However, water in excess of that capable of enter- 
ing into the composition of the crystals destroys them, either by solution 
or decomposition. 

The crystals formed from calcium oxide are in all respects similar to 

^'^Richter, W., Thonindustrie Ztg. (1903), 1863. This method only applicable 
to fresh, dry clinker. Am. Journ. 8ci. (1906), 172, 266. 
^^ Log. cit. 
^"^ Log. cit. 



372 



REIBLING AND REYES. 



those from the hydroxide, but in this case more water is required to 
complete the crystallization because the oxide must first hydrate and in 
so doing absorbs about one-third of its own weight of water. 

These facts were made clear by a method which is illustrated in 
figure 1. 

3 ^ ^ - 




CA 



CAO 

Fig. 1. — Sealed tubes showing the effect of water on the formation of calcium hydroxide- 
phenol crystals. 

Five test tubes were taken, each of which contained 0.25 gram of pure, dry 
calcium oxide or hydroxide, 30 cubic centimeters of absolute chloroform, 1.5 gram 
of phenol and the amount of water specified in the illustration. The tubes were 
sealed, shaken for eight hours, and then their contents allowed to settle. Calcium 
oxide and hydroxide have a higher specific gravity than chloroform. In tubes 1 
and 2, where no water was present, they were not acted upon, and have settled 
to the bottom of the liquid with their original bulk imchanged. On the other 
• hand, the crystals of calcium hydroxide-phenol are specifically much lighter than 
chloroform. In tube 3 with two drops of water there was incomplete reaction 
and the mixture does not show a sharp line of separation. In tubes 4 and 5 with 
four and six drops of water respectively, the change has practically reached com- 



PROPERTIES OP PORTLAND CEMENT. 373 

pletion and although the crystals may not be pure, still all of the solid has 
risen to the surface and a sharp line of demarcation is visible. A microscopic 
examination of the crystals formed in tubes 4 and 5 showed that they resembled 
each other in all respects. They were similar to tho^ shown by figure 2. 

A more satisfactory and accurate method for tracing the gradual slak- 
ing of lime and its ultimate conversion into the crystals under considera- 
tion, is to allow hydrogen gas saturated with water to bubble through 
chloroform or nitrobenzol containing a slight excess of phenol over and 
above the amount necessary to bring the calcium oxide into complete 
reaction. A microscopic examination of calcium oxide treated in this 
manner shows that the reactions proceed very slowly. 

At first, a gradual hydration of the oxide takes place, which can be noted 
because the lime when hydrated becomes visible in polarized light through crossed 
nicols. After the gas has bubbled through the solution for many h6urs, small, 
slender crystals begin to appear. These continue to grow in size and number, 
and the loss by ignition of filtered samples gradually approaxjhes a maximum. 
Eventually, the crystals resemble those which are developed by a similar process 
from the hydroxide and if the saturated gas is allowed to continue to bubble 
through the solution, the loss by ignition of filtered samples decreases until 
finally they dissolve in the solution. If the latter is then dehydrated, the crystals 
again appear. It requires very much less saturated gas to produce a maximum 
yield of crystals from the hydroxide than from an equivalent amount of oxide. 

The crystals have also been separated by us from their solution in an 
excess of phenol and water. Water gradually was added to the tubes 
containing the oxide or hydroxide^ phenol, and chloroform or nitro-benzol, 
until the crystals formed had dissolved. The resulting solution was 
filtered, evaporated over sulphuric acid to crystallization in a vacuum 
desiccator, the crystals were then washed with chloroform and dried in 
vacuo. 

Unfortunately, no solvent for recrystallization has been found, and 
therefore it has been impossible to obtain exactly concordant numbers on 
analysis. The purest crystals obtained were found to contain from 20.5 
to 23.8 per cent of calcium oxide. 

The analyses of the purest crystals as well as the conditions necessary 
for their synthetic preparation, indicate that they have the following 
composition: C8i{0B.)^, {C^'EfiB.)^,B.fi.^^ 

A study of the behavior on the microscope slide of calcium oxide and 
hydroxide in the presence of phenol containing varying amounts of water 
confirms these conclusions and gives a method for the positive identifica- 

^' For example, in one instance 0.3280 gram of substance gave 0.0778 gram 
of calcium oxide and 0.1592 gram of water. A microscopic examination of these 
crystals showed that they contained a small amount of impiirities. If we con- 
sider these impurities as calcium carbonate and the above formula as the com- 
position of the pure crystals the theoretical yield of water should have been 
0.1676 gram, which differs by only 0.0020 gram from that actually obtained. 



374 



REIBLING AND REYES. 




Fig. 



2. — Characteristic crystals obtained by the 
scopic test from calcium hydroxide. 



tion of free lime, sintered, non-sintered and slaked^ in Portland cement 
and like materials. 

The amounts of oxide and 
hydroxide used on the slide 
were varied from a mere 
trace to 3 milligrams; and 
only those crystals were 
compared which formed in 
the thin sections under the 
cover glass where the par- 
ticles were well distributed. 
The reasons for so doing are 
obvious. The cover glass on 
the slide keeps the solution 
motionless. Tlierefore, each 
particle of oxide or hy- 
droxide is confined more or 
less to that part of the 
solution which is in contact 
with it. Either substance 
tends to dehydrate and de- 
phenolate the surrounding 
solution and thus to retard 
or prevent the further 
growth of the crystals. 
This is shown by the fact 
that in solutions containing 
less than l.G per cent of 
water, the average length 
of crystals formed from 
isolated particles is greater 
than from those which are 
massed together. 

Theoretically^ we would 
expect to find a radical 
difference between the mi- 
croscopic results obtained 
from the treatment of the 
oxide and the hydroxide 
with the test solution ^^ 
advocated by White and 
in reality this is the case. 
If we calculate from the 
formula of the crystals it 
is evident that one drop (50 milligrams) of White's solution would contain 
enough water to react with 0.92 milligram of calcium hydroxide and 
produce 3.59 milligrams of the crystals. On the other hand, one drop 




Fig. 3.- 



-Characteristic crystals obtained by the micro- 
scopic test from calcium oxide. 



^•Five grams of phenol -f 5 cubic centimeters of nitrobenzol + 1 drop (0.05 
gram) of water. 



PROPERTIES OF PORTLAND CEMENT. 



375 



of this solution would contain only enough water completely to hydrate 
about 0.67 milligram of calcium oxide. This difference is apparent in 
the microphotographs shown by figures 2 and 3 and is also evident from 
the data recorded in Table I. 

Table I. — The effect of water upon the length of calcium hydroxide-phenol crystals 
formed on the microscopic slide. 



Number of 
drops of 

water 
added to 
the anhy- 
drous 
phenol 
solution. 



0.0* 
0.0 
1.0 
4.0 
7.0 



Percent- 
age of 
water in 
the test 
solution. 



0.0 

(^) 
0.40 
1.60 
2.80 



Length of crystals, in milli- 
meters, produced in the 
thin sections of slides. 



From calcium 
hydroxide 

[Ca(0H)2]. 



From calcium 
oxide (Cao). 



No crystals. 



0. 06-0. 12 
0.06-0.12 
0. 06-0. 27 
0.27-0. 60 -f 



0. 015-0. 03 
0.015-0.03 
0.015-0.03 
0.025-0.06 



» The anhydrous reagents were prevented from absorbing water from the air by ringing 
the cover glass with Canada balsam. The same materials developed crystals in about 
fifteen minutes if the cover glass was not so inclosed. 

b Trace absorbed from air. 

Table I shows that crystals from calcium oxide do not grow in a motionless 
phenol solution containing less than 1.6 per cent of water, to a greater length 
than 0.03 millimeter. On the other hand, the presence of a mere trace of that 
reagent was sufficient to cause them to grow from calcium hydroxide in lengths of 
from 0.06 to 0.12 millimeter. We have not been able under a tight cover-glass 
inclosed in balsam to produce crystals from calcium oxide longer than 0.270 
millimeter except by the use of phenol saturated with warm water. This same 
solution dissolves the crystals if the hydroxide is used; and water saturated with 
phenol dissolves those formed from both oxide and hydroxide. 

In solutions containing less than 1.5 per cent of water, the presence of a con- 
siderable amount of oxide tends slightly to reduce the length to which crystals 
produced from the hydroxide would otherwise grow. 

All of the crystals intermingle freely and retain their characteristic 
appearance, thus further proving that the final crystalline compounds 
obtained from both the oxide and hydroxide are chemically identical. 

The difference in the size of the crystals which are formed in slightly 
hydrated solutions under the conditions imposed by the microscopic tes% 
is due entirely to the fact that the oxide must hydrate before it is able to 
react with phenol and more water to develop calcium hydroxide-phenol 
crystals. This tendency toward hydration is so great that phenolization 
only occurs to a slight extent until water in excess of one-third of the 
weight of the oxide is available. 

These results make it highly probable that the long, radiating crystals 
which are shown in White's photomicrograph ^^ of a commercial cement, 

^Loc. cit, (figure 1). 



376 REIBLING AND REYES. 

were produced from slaked lime and not from calcium oxide as he stated. 
White emphasizes the time factor which enters into the reaction. The 
time required for the formation of these crystals depends primarily upon 
the amount of water which is available at all stages of their development. 
We found that on increasing the percentage of water in the test solution^ 
the crystals formed more and more quickly and finally, 9 drops of water 
in 30 cubic centimeters of the solution produced them almost immediately. 

Another habit of formation of calcium hydroxide phenol ciystals 
remains to be considered. White states that "the crystals formed from 
lime which has been fused in the electric arc * * * appear as plumes 
or feathery petals which in favorable cases give the group somewhat the 
appearance of a chrysanthemum.^^ ^^ However, the development of plume- 
shaped groups of crystals does not necessarily indicate fused limed. 

Lime melts at an extremely high temperature; so high, in fact, that 
Arthur L. Day and E. S. Shepherd ^^ found it impossible to make a 
satisfactory determination of its melting point. They state that "lime 
can be fused in the electric arc under favorable conditions.^^ We were 
unable to prepare homogeneous, fused lime in the electric arc of the ap- 
paratus at our disposal, but the sintered mass which we obtained produced 
plume-like aggregates of crystals in abundance when treated with Whitens 
reagent. Similar forms were also secured by us from pure marble which 
had been heated in a porcelain crucible on the forge and from under- 
burned Portland cement clinker the temperature of burning of which was 
known to be much too low to admit the possibility of the presence of fused 
lime. 

White also found plume-shaped crystals in Portland cement ^^ burned 
by E. D. Campbell at a temperature below 1,500 degrees. 

We are inclined to question the presence of fused, free lime even in well- 
burned Portland cement clinker. The highest temperature of burning 
is seldom above 1,700 degrees and in the rotary process especially, it is 
not probable that the free lime would absorb enough heat units in a 
sufficiently short time to cause it to melt, except perhaps at local, very 
minute points, if the effect were similar to that found in the mantle of a 
Welsbach burner. 

However, since it has been found impossible to produce plume-shaped 
crystals from lime which has only been heated in the ordinary blast lamp, 
it is evident that it must at least be sintered before the habit of the 
crystals under consideration can be observed after the addition of phenol 
and water. 

The reason for the . formation of crystals in plume-shaped aggregates 
becomes evident by a study of the method of their formation. 

2i hoc. dt. 

^Journ. Am. Ghem, 8oc. (1909), 28, 1089. 

^ Log. cit, (figure 3). 



PROPERTIES OF PORTLAND CEMENT. 377 

The plumes are composed of a more or less compact mass of fine, slender 
crystals, each individual of which is similar in size and appearance to those 
developed from fine particles of nonsintered calcium oxide. They develop from 
certain points on the surfaces of particles of considerable size which form their 
nuclei and may be desintegrated and their component parts made more apparent 
by moving the cover glass after they have formed. The first needles formed are 
pushed out into the surrounding liquid by those which follow. The cover glass 
forces the liberated crystals to spread in the direction of least resistance and in 
so doing they cluster into flat groups. The fine, slender crystals which compose 
these aggregates give to them their characteristic feathery appearance. 

There appear to be points or spots on the surfaces of sintered or fused 
particles of lime which are most easily attacked by the test solution. We 
can assume that a fracture, abrasion, or point less heated than others would 
more easily be hydrated than smoother surfaces, and that such places would be 
the first to be attacked by the water in the test solution. The hydration of 
the lime causes it to swell and to push the crystals into the surrounding liquid. 

^ Calcium hydroxide when it is formed by small quantities of water 
acting on sintered particles of lime is crystalline and can be seen in 
polarized light through crossed nicols, the whole cluster thus becoming 
visible. The crystals on the outer edges are acted upon by phenol more 
rapidly than the 
others and thus 
feathery fringes, 
which are character- 
istic of the plumes, 
are formed. This is 
shown by the follow- 
ing experiment: 

Lime which had been 
heated in the electric 
arc was allowed to react 
with a slight excess of 
water, thinly spread over 
the surface of a micro- 
scopic slide. This slide 
was dried in a desic- 
cator without exposure 
to carbon dioxide and 
then carefully covered 
with nitrobenzol and a 

cover glass. It was 

,, . J . Fig. 4. — Crystalline calcium hydroxide obtained from sintered 

then examined in po- Uj^^ 

larized light through 

crossed nicols. The result is shown by figure 4, which is a photomicrograph of the 

crystals formed in this manner. The skeleton outline of plume-like forms, which 

subsequently become so plain after adding phenol and water, can readily be seen. 

Fused lime will act in a manner similar to the sintered body, with the 
difference that the crystals would develop much more slowly. Therefore 
the appearance of plume-like crystals is a positive indication of rather 




378 



REIBLING AND REYES. 



coarse particles of lime which have been calcined at a high enough 
temperature to cause at least the sintering of their surfaces. 

So much for the appearance and methods of producing the crystals 
of calcium hydroxide-phenol as they are formed with lime only. Those 
secured by treating Portland cement are similar to the crystals formed 
either from pure calcium oxide or hydroxide. 

Underburned, nonseasoned commercial cement clinkers, pulverized and with or 
without the addition of plaster, develop crystals with the test solution which 
are identical in appearance with those which we have obtained from calcium 
oxide. Figures 5 and 6 are photomicrographs obtained respectively, from pure 
calcium oxide and from nonsintered lime from Portland cement and figures 
7 and 8 from lime heated in the electric arc and sintered lime in cement. A 
comparison will make the fact clear. 




Fig. 5. — Characteristic crystals obtained by 
the microscopic test from cal- 
cium oxide ignited at a red heat. 



Fig. 6. — Cbaracteristic crystals obtained by 
the microscopic test from non- 
sintered lime in Portland cement. 




Fig. 7. — Characteristic plume-like forma- 
tions of crystals obtained from 
sintered lime. 



Fig. 8. — Characteristic plume-like forma- 
tions of crystals obtained from 
sintered lime in Portland cement. 



PROPERTIES OF PORTLAND CEMENT. 379 

The same cements shown by figures 6 and 8, when seasoned in moist 
air give crystals (see figure 9) which correspond exactly to those secured 
with calcium hydroxide (see figure 2). 

An increase in the percentage of water in the test solution also has 
the effect of producing crystals similar to those shown in the last figure. 

It might be argued that 
in cement, certain silicates 
or aluminates of calcium 
might be decomposed by the 
test solution to give crystals 
identical with those pro- 
duced from free lime. Apart 
from the fact that such a 
discussion would be purely 
academic, were compounds to 
exist so easily decomposed 
by phenol, yet it may be 
stated that perfectly sinter- 
ed, aerated and nonaerated 
clinkers, ground to pass a 
200-mesh sieve, fail to give 
White's reaction; so that it 
appears not to be character- 
istic of lime which has 
entered into combination at 

Fig. 9.— Characteristic crystals obtained from slaked lime ^^S^^^ temperatures with 
in Portland cement. other substances in cement. 

Therefore, White's test offers a positive means for the identification of 
sintered, nonsintered and hydrated free lime in such materials as Port- 
land cement; however, only an experienced operator who thoroughly 
understands the conditions of the reaction in which three different 
characteristic habits of the same crystal can be produced under given 
circumstances can do this successfully. 

It is often extremely difficult, if not impossible, to estimate accurately 
the relative quantities of the different kinds of free lime which are 
present, but the conclusions to be drawn can often be rendered more 
certain by examining several slides in test solutions containing different 
percentages of water. A study of the material retained on sieves of 
different sizes may also give valuable information. The test solution is 
hygroscopic and must be protected from moisture. 

APPLICATIONS OF THE TEST FOR FREE LIME. 

It has been stated that "cement which comes from the kiln, containing 
more than a trace of free lime, will not pass the boiling test until the 
free lime has been removed.^^ However, in spite of the fact that all of 
the cements recorded in Table A were perfectly sound, they each showed 




380 REIBLING AND REYES. 

more than a trace of free lime when the test was made. Short and long 
needle-like crystals appeared in all the slides, and plumes in many. 
Cements which disintegrate were also carefully examined for comparative 
purposes. They gave a relatively greater number of crystals than the 
majority of the sound commercial cements, but the distinction is not 
absolute, because the cements marked ^'B/' ^'0/^ "J/^ and "M/' which 
belong to the class, contained enough free lime to render them indis- 
tinguishable from unsound cements. 

We prepared and sealed in test tubes duplicate samples of eight com- 
mercial cements and four ground by us from the clinker. One set was 
reserved for future reference and the other sent to Alfred 0. White, who 
also identified free lime in all of them and classified six as '^^poor^^ and 
two as ^^passable.^^ He failed to detect a sound and unsound cement of 
the same brand, classifying them both as ^^poor.^^ (See Table YIII.) 

One of us, while at a large manufacturing plant, examined cements 
which could even withstand the severe Faija boiling test and these also 
frequently gave an abundance of needles and plume-like crystals when 
they were subjected to microscopic study. 

The limitations of the test for soundness as an indicator of proper 
burning were made manifest in the previous publication ^^ from this 
laboratory. To judge by the results given above, it is evident that lime 
may exist in considerable quantity in some sound Portland cements. 
Therefore we endeavored' to ascertain to what extent the soundness and 
other properties of Portland cement were affected by various amounts of 
free lime under different conditions of burning, grinding, and seasoning. 

The following list of observations (Table II), showing the effect of the tem- 
perature of burning upon the habit of the crystals of calcium hydroxide-phenol, 
were made possible by an accident to the gearing of a 28.5 meter rotary kiln 
which enabled one of us to inspect its interior and to collect samples of material 
throughout its length: These samples were immediately sealed and at a later 
time examined by means of the phenol test. The shortness of the clinkering 
zone of this kiln is characteristic of the rotary process. It began at about 5 
meters and ejided at about 1.5 meters from the hot end of the furnace. 

-'This Journal, Sec. A (1908), 3, 137-185. 



PROPERTIES OF PORTLAND CEMENT. 



381 



Table II. — The effects of increasing temperature on rfiw material {rotary kiln 
process with limestone and clay ) . 



Sam- 
ple 

num- 
ber. 


Distance 
from the 
hot end of 
furnace 
at which 


Color of 
material. 


Physical 
condition. 


Microscopic test with phenol 
(duplicate specimens). 


sample 
was taken. 










Meters. 








7 


16.4 


Pinkish-gray 


Not sintered. As 


No crystals developed within 2 hours. 






Pink color due to 


fine as the raw 


The material contains no calcined 






the presence of 


meal. 


limestone. 






free ironoxide. 






11 


10.0 


Light yellow 


Slightly sintered 


In 5 minutes the slide showed a mass 






Change in color 


and gritty. 


of radiating clusters and single short 






due to less free 


White particles 


needles. A few plumes also appeared. 






iron-oxide and 


and a few 


The material contains a considerable 






to the forma- 


small porous 


amount of calcined limestone in the 






tion of a green- 


lumps (pea 


form of free, nonslntered lime. A 






ish-brown sub- 


size). 


little free lime has sintered. 






stance. 






13 


7.5 


Light to dark 


Slightly sintered 


An almost immediate and abundant 






brown. Spot- 


into small, soft, 


formation of plumes and some clus- • 






ted with white 


porous lumps. 


ters of short radiating needles ap- 






specks. 




peared. Most of the free lime has 
become sintered. 


15 


6.2 


Brown to black. 


Brittle but por- 


An almost immediate formation of 






Spotted with 


ous and soft. 


plumes in great numbers, no radiat- 






white specks. 


Easy to grind. 
Sintered into 
large and small 
lumps. 


ing nor separate needles. All of the 
free lime is sintered. 


18 


4.6 


Steel-gray to 


Very brittle but 


Plumes in abundance but not as many 






black with 


still porous. 


as sample No. 15. The percentage 






white specks. 


Hard to grind. 
Large and 
small lumps. 


of free lime has decreased. 


22 


Outlet. 


Steel-gray and 


Rounded lumps. 


Fair amount of plumes. Some single 






still slightly 


Well sintered, 


clinkers gave a negative test for free 






spotted with 


dense and hard. 


lime. The material contains only a 






white specks. 


Very hard to 
grind. 


small amount of free lime. 



382 



REIBLING AND REYES. 



Samples of well-burned and underburned stationary kiln clinker fresh 
from the kilns, were also collected and then subjected to the examination 
recorded in Table III. 



Table III. — Effects of different temperatures of burning on rcuw-material {station- 
ary kiln process with limestone and clay) . 



Brand 
of ce- 
ment. 


Kind of clinker. 


Fineness 
through— 


Microscopic test (duplicate 
specimens). 


Soundness 

(5-hour 
steam test). 


200 
mesh. 


100 
mesh. 





Well-burned clink- 
er. 


93.6 


100 


An almost immediate forma- 
tion of plumes and feathery 
rosettes. No needles until 
after 10 minutes, when a few 
formed. 


Slightly disinte- 
grated. 






U n d erbwrn ed 
clinker. 

Selected, thorough- 
ly fused clinker 


89.4 
83.8 


100 
99.6 


Mass of radiating clusters and 

single needles and plumes. 
No crystals in 1 hour 


Slightly disinte- 
grated. 
Perfectly sound. 






which Bhowed no 












white specks 
when fractured. 










L 


Well-burned clink- 


93.6 


100 


An almost immediate formation 


Not disintegrat- 




er. 






of plumes. 


ed but scales 
on bottom sur- 


L 


U nde r burned 
clinker. 


90.8 


100 


A mass of radiating clusters and 
single needles and plumes. 


face, and off of 
the plate. 
Completely dis- 
integra ted 
and badly 


L 


Selected, thorough- 
ly fused clinker 


100 


100 


No crystals in 1 hour 


swollen. 
Perfectly sound. 






showing homoge- 












neous composi- 












tion and no white 












specks. 











The data recorded in Tables II and III show that in the manufacture 
of Portland cement, as the temperature is increased the nonsintered 
calcium oxide gradually is converted into the sintered material, having 
different physical properties. This naturally is what might be expected ; 
but this conversion may occur at temperatures far below those necessary 
for the proper burning of Portland cement clinker. Therefore, under- 
burned cement may contain both sintered and nonsintered lime, but the 



PROPERTIES OF PORTLAND CEMENT. 



383 



presence of the latter in unadulterated Portland cement is a positive 
proof of underbuming. 

Free, sintered lime may exist in dense, thoroughly clinkered cement 
material if the conditions essential for yielding a completely homogeneous 
product, such as fineness, duration of exposure to high temperature, 
proper chemical composition, etc., have not been complied with. On 
the other hand, Portland cement clinker may be burned perfectly so as 
neither to contain free lime nor have lime liberated in the ordinary process 
of cooling and grinding. 

The data recorded in Tables TV, V, and VI show the effects of the 
atmosphere on sintered, nonsintered and slaked lime in Portland cement 
under different conditions of aeration. 

Conditions of the experiment recorded in Table IV. — Five grains of the cement 
(Brand 0, Table III) which for the greater part gave plumes on microscopic 
examination, were exposed to the atmosphere in a 6-millimeter test tube. At 
regularly increasing intervals of from thirty minutes to one day, the test tube 
was corked, its contents thoroughly shaken, and a test made from the resulting 
mixture. 

Table IV. — Microscopic tests of an aerating cement containing uncombined 

sintered lime. 



Sam- 
ple 

num- 
ber. 



Not aerated. 



Aerated 2 hours __. 
Aerated 8 hours __, 



Aerated 32 hours __ 

Aerated 80 hours— 
Aerated 272 hours. 
99667 2 



Microscopic test (duplicate samples). 



An almost immediate formation of plumes, some of which were 
fully developed in a few minutes.* Others developed more slowly, 
requiring as long as 15 or 20 minutes to complete the reactioii. 
At first no individual needles appeared, but after 10 minutes a 
few short radiating clusters were discovered. 

Similar to the above, except that a careful comparison of many 
samples indicated the presence of long needles. 

Many plumes develop almost immediately. In some instances their 
edges were first surrounded with rather long radiating needles. 
These were soon pushed aside and so formed an outer smooth edge 
to the plumes which could not be distinguished from the rest of the 
crystalline formation. These slides show a marked increase in the 
relative amount of needles, most of which radiate from central 
nuclei in the form of rosettes. 

A mass of rosettes and clusters formed of long, needle-shaped crys- 
tals is shown, A few part feather and part needle rosettes are 
formed. 

No plumes nor feathery crystals appear. The slide shows clusters 
of long needles and rosettes in abundance. 

No crystals whatever appear within 1 hour. 



384 



REIBLING AND REYES. 




Pig. 10 (magnification X 180). — Photomicrograph of calcium hydroxide-phenol crystals 
of the unaerated cement recorded in Table IV. It shows the typical plume-like 
habit of the crystals. Crystalline formations of this character are characteristic 
of coarse particles of sintered lime which is free from even partial hydration. 




Fig. 11 (magnification X 180). — Photomicrograph of the crystals from the same cement 
after it had been aerated for eighty hours (see Table IV). It shows the rosettes 
and the long slender needles which are characteristic of the hydroxide. This figure 
represents the final stage In the process of hydration of the sintered lime, 1. e., 
the total disappearance of plume-like crystals. 



PROPERTIES OF PORTLAND CEMENT. 



385 



The intermediate stages in the process of hydration can fairly well be 
traced without illustrations by the data recorded in Table IV. These 
include the mixed formation of long crystals on the edges of the plumes, 
and the production partly of needles and partly of feathery rosettes, 
which show that sintered lime may hydrate so slowly by mere exposure 
to the atmosphere that the action takes place essentially on the outer, 
exposed parts of the particles and only gradually penetrates into their 
interior. 

The negative test for free lime at the close of the series recorded in 
Table IV is due to the final, complete conversion of at least the surfaces 
of the hydrated lime into carbonate. 

Conditions of the experiment recorded in Table Y, — Well-burned and under- 
burned nonaerated Portland cements were used in this experiment. Each was 
spread in a layer about 1 millimeter thick on the bottom of a crystallizing dish 
and exposed to the air. 

Table V. — Comparative microscopic tests of two cements containing free lime 

aerated in thin layers. 



Time. 


Microscopic test (duplicate samples). 


Well-burned clinker (Brand 0, 
Table III). 


Underburned clinker (Brand 0, 
Table III). 


Not aerated 


An almost immediate formation 
of plumes. A few radiating 
clusters of needles after 10 
minutes. 

No needles formed in 20 minutes. 
After 30 minutes many plumes. 
No needles except on the 
edges of some of these. 

No needles formed in 30 minutes. 
In 1 hour several plumes and 
feathery rosettes had appeared, 
but no needles. 

No crystals in 30 minutes. In 1 
hour only 10 feathery rosettes 
and 2 plumes appeared in 
2 specimens. 

No crystals in 30 minutes. In 1 
hour only 2 plume-like aggre- 
gates could be found. 

No crystals of any kind appeared 
inli hours. 


In 6 minutes a mass of rosettes, 
clusters and needles, some 
plumes. 

Ill 10 minutes no crystals. In 20 
minutes many rosettes of 
needles and some plumes. In 
1 hour about the same. 

In 20 minutes no crystals.. In 
1 hour plumes and feathery 
rosettes but no single needles. 

In 30 minutes no crystals. In 
1 hour a few feathery rosettes 
and plumes. 

In 1 hour no crystals had 
formed. 

No crystals of any kind appear- 
ed in U hours. 


Aerated 1 day; slightly 
caked. 

Aerated 2 days 


Aerated 5 days 


Aerated 7 days 


Aerated 8 davs 





The results given in Table Y differ very much from those of Table IV. 
It must be understood that the penetration of air into a mass of ground 
cement is limited approximately to a thin outer layer. (See Tables X 
and XI.) Almost every particle of the cement referred to in Table V 
was fully exposed to the atmosphere at all times, while the material 



386 



REIBLING AND REYES. 



providing data for -Table IV was arranged so that only an upper layer 
was fully exposed at any given time. 

The absence of individual needles and the presence of aggregates form- 
ing plumes at times during the^ course of aeration supports Whitens state- 
ment that ^^experiment on the aeration of cement in thin layers in the 
laboratory indicate that the conversion into carbonate goes practically 
as fast as hydration/^ It will be noted that as the length of the time 
of aeration increased, that necessaiy for the formation of calcium hy- 
droxide-phenol crystals also increased. Since sintered calcium oxide 
may hydrate very slowly and since the slaked lime changes to carbonate 
very rapidly, it seems reasonable to suppose that particles of sintered 
lime gradually became coated with a film of calcium carbonate which 
prevented the immediate action of the testing solution. It also is very 
probable that this outer coating of carbonate might become so dense as 
finally to prevent hydration and combination with phenol in the inner 
portions of tl\e particles. The work recorded in Table VI was therefore 
undertaken to ascertain if this conclusion is correct. 

Conditions of the experiment recorded in Table YI. — 'The cements recorded 
in Table V, which had been aerated until no crystals could be formed on the 
microscope slide, were each reground in an agate mortar and again subjected to 
the test for free lime. 

Table VI. — Microscopic tests of reground, aerated cements. 



Description of Port- 
land cement. 



Well-burned clinker 
aerated 8 days 
(Brand O, Table 
III). 

The same reground.. 



Underbumed clink- 
er aerated 8 days 

• (Brand O, Table 
IV). 

The same reground. , 



Microscopic test (duplicate samples). 



No crystals formed in 1 hour.. 



An almost immediate formation of crystals. In 5 min- 
utes the slide presented a network of clusters, long and 
short, single and radiating needles and of plumes. 

No crystals formed in 1 hour i_. 



Indications of plumes in 2 minutes. In 5 minutes the 
slide presented a network of long and short needles 
and well-formed plumes. This underburned cement 
shows a greater quantity of plumes and needles than 
the reground, well -burned material. 



Soundness 
(5-hour steam 

test). 



Sound. 



Do. 



Do. 



Slightly dis- 
integrated. 



These results make it evident that free lime may continue to be 
present in considerable quantity even in thoroughly aerated, finely ground 
cement, although such cement may have ceased to form calcium hy- 
droxide-phenol crystals. The significance of this fact will become more 
apparent elsewhere. 



PROPERTIES OF PORTLAND CEMENT. 387 

Selected, thoroughly sintered, unaerated cement clinkers (Brands 
and L, Table III), which gave a negative test for free lime, were treated 
in a manner similar to that recorded in Tables IV, V, and VI. In no 
case could free lime be detected. 

Conditions of the experiment recorded in Table VII. — ^We have endeavored 
to establish a definite relationship between the microscopic evidence of the 
presence of free lime and the physical condition of the cement as evidenced by 
the tests for soundness. Aeration tends to remove unsoundness and a simulta- 
neous study of the influence of the air upon the free lime and upon the soundness 
of the cement was resorted to in an effort to compile conclusive data. The cement 
used for this experiment was prepared by grinding about 10 kilograms each of 
so-called "good" and "bad" clinker in a ball-mill, until no residue was left 
on a 100-mesh sieve. The cement was somewhat exposed during the process of 
grinding. Two per cent of plaster of Paris had been added to the clinker. 
The cement was aSrated in 300-cubic centimeter wide-mouthed bottles. Each 
bottle contained 110 grams of cement which was mixed thol-oughly each day. 
Table VII records the results. (See pp. 388, 389.) 

That the effects of free lime upon soundness are influenced by the 
cohesive properties of the cement becomes apparent from the results 
recorded in Table VII; and it is evident that the cause of unsoundness 
was more difficult to eliminate from the underbumed material than from 
that which was well burned. The microscopic examination of the cements 
which had been aerated during two days, failed to show the differences 
which should have appeared if the physical tests alone were taken into 
consideration. We attribute the discrepancy to the presence of more and 
stronger cementive material in the better burned product, our universal 
experience being that underbumed cements at any stage of aeration or 
seasoning do not develop the high early strength characteristic of well- 
burned Portland cement. 

These experiments all indicate that the usual cause of unsoundness in 
cements is unslaked free lime, and that commercial Portland cements 
may develop perfect soundness and still contain a considerable quantity 
of free lime. The limitations of the soundness test as a test for free 
lime or for underbuming are self-evident, but White's test is absolute. 

These results as well as many others which we have obtained also seem 
to prove that from the microscopic evidence alone the behavior of all 
cements when subjected to the soundness test can n^t be predicted. It 
is doubtful, if we consider the character of the physical test, if even a 
quantitative measurement of sintered, nonsintered, and slaked lime 
would prove entirely satisfactory in this respect. Unsoundness in Port- 
land cement becomes manifest to the eye only when the force or farces 
which operate to cause disruption are sufficiently great to overcome the 
cohesion between particles. Therefore, the development of cohesive 
forces, the ^^speed of slaking,'^ fineness, the temperature and amount of 
water used in gauging, and the effect of impurities and retarders, must 
all be taken into consideration; and consequently the test for aoundness 



388 



EEIBLING AND REYES. 



I 






CO 



« 93 23 

»^ ^ o s 






>• V. 



'S s 



S 2^ <n 



<u o 

q -M 00 



^ ^ 



5*1 



J> ft! ^ 
O O - 



03 03 



-2 ft :S ? 



3 a 
a ^ 

" (15 



S ^• 



9 " 9 

$ 't ^ : 

t* '3 ta " 

s a :: 

A XT, O 

« ^ I 



0^ 



«j a 



orj 


ft 






S 




^ 


0) 


S 


a 


(1) 


o 


'O 


a 


a> 



•^ S b 



S ? ^ 

S § ^ 
fl ft^ 

>» M.H S 

C O +j 

a fl 2 
•• I "^ 

03 O " 

g "^ § 



C H 0) 

.2 '5 -p 



. a. 
a ^ 



1-^ 

.2 ^ 

a ^ 



a 



SI? 



^ V .M 

Sag 
< §§ s 

a) «j 

1 ^ § 
•9 2 a 
a a I 

"Soft 



a 



Vaqranu amog 



<5 






"S 

« 



•«-l <» 






•2 5 a 
a S ^ 



5 a s 
S2 go 

:aj .w ^ 



-o »« 


o 


CO lO 




'd TS 


"O 


^ ^ 


-2 


g g 


2 


:(X) :a> 


:a> 


^ <3 


< 


'"f tn 


o 



PROPERTIES OP PORTLAND CEMENT. 



389 




•S S 



s a 



t^ 




0) 


-o 


a 




^ 


fl 


T) 


0) 


0) 


6c 



S .2 



•i^ t 



fi « ^ 



5^^ I 

O 2 03 g 

a i 



j5 'O 

a 52 



o 




£ 


11 


^ 


^ 
i 








ri 


bo X 


oJ 


9i 






W 




!5: 






I ^ ^ 








MJ 






0) 0) 








a> 






5 s "^ 














o 






















fl 






and 
same 
lume 








1 






lumes 
utes: 
no p 








03 

o 

03 






a fl ^H 








ton 






|a| 








a 
o 














o 






•^ - fl 








OT 






- C *-H 








0) 






03 t-( 














TS 














.5 d S 








o 






^ 0) o 








•"• 






fall 
pear 
mat] 














o ft t 














$ -•2 


S 












needl 

not 

r-like 


1 
p 






l; 
^ 








390 REIBLING AND REYES. 

as indicating the presence of free lime is relatively crude as compared 
with the microscopic study of calcium hydroxide-phenol crystals. 

The following considerations will give a clearer understanding of the 
effect of free lime upon the soundness of cements. 

Slaked lime has little or no effect upon soundness. As much as 50 
per cent can be added to some cements without causing them to check 
or disintegrate. 

The speed of slaking is increased by the fineness of gi'inding. The 
experienced operator in preparing the microscopic slides for the tests for 
free lime can obtain a very good conception of the fineness of the material 
during the process, because a microscopic measurement of the sizes of 
the particles gives definite information in this respect. 

Nonsintered lime slakes in an excess of water at a low temperature. 
The lime will obviously have no effect per se on the soundness, pro- 
vided it slakes during the mixing with water, or before setting has 
begun. Portland cement manipulations can not guarantee an excess of 
water, or the time necessary for the complete slaking of nonsintered lime 
before setting has developed. If the operation is conducted under the 
American Society methods of testing, then a considerable proportion of 
nonsintered lime will have had ample opportunity to slake in a ^^normal 
consistency^^ mixture. Therefore, nonsintered lime must ordinarily have 
been present in considerable quantity to cause disruption in cements so 
tested. On the other hand, nonsintered lime is not present in well-burned 
cement. Its presence is a positive evidence of underburned, mixed, or 
adulterated material and therefore the soundness of cements containing 
a considerable quantity of nonsintered lime is very uncertain, as we 
can not rely upon the early development of strong cohesive properties to 
offset any force or forces which may operate to cause disruption. 

"Fused lime slakes very slowly in cold water. Hot water slakes it 
more rapidly.^^^^ As the degree of incipient fusion approaches that of 
perfect fusion the lime will become more and more inactive in cold water. 
Therefore, some cements have remained sound in cold water even for 
months, and yet disintegrated when placed afterwards in boiling water,^^ 
the disrupting force of expansion developing so slowly in the cold that 
the growing forces of cohesion keep ahead of it. 

Sintered lime may exist both in under- and well-burned cement, but 
the sintered lime in hard-burned cements probably slakes less readily at 
normal temperatures. 

Experiments in this laboratory indicate that certain cements so 
regulated by retarders as to set slowly may remain sound, in the steam 

^"Day and Allen, Journ. Am. Chew,. Soc. (1909), 28, 1094. 
2« Loc. cit. 



PROPERTIES OF PORTLAND CEMENT. 391 

test and yet develop unsoundness if the set is quickened by the use of 
more or less plaster; so much sintered lime remaining unslaked in the 
quick-setting cement as to cause unsoundness in the accelerated test. 
For the same reason plastered^ nonseasoned cements have been en- 
countered which have developed unsoundness in the steam test, but which 
remained sound when no plaster had been added.^^ 

These facts will more fully be discussed in another part of this paper. 
We mention them here in order to emphasize the limitations of the 
soundness test as a test for free lime. 

However, taking all factors into consideration it seems as if the follow- 
ing conclusions would enable us more accurately to interpret the relation- 
ship between free lime and soundness.^® 

1. Slaked lime will not cause unsoundness in the soundness test. 

2. Nonsintered lime must be present in quantity to cause unsoundness ; 
and if it is .so present, the disruption is likely to occur in water and air, 
as well as in the steam pats. 

3. If both sintered and nonsintered lime are present, the pat may be 
sound in air and in water, but it is very liable to disintegrate during the 
hot tests. 

4. Fused or sintered lime, in the absence of non-sintered lime, in- 
dicates a hard-burned cement which in all probability contains an 
abundance of material having strong cohesive properties. The effect of 
such lime upon the soundness is therefore a quantitative consideration. 
Sintered lime, unless present in large amounts, will not cause disrup- 
tion in air and water pats, but even a fair amount is liable to cause 
unsoundness in accelerated tests. The effect of a small amount is not 
liable to become apparent to the eye. 

5. Fineness assists the soundness. 

Conditions of the experiment recorded in Table VIII. — These conclusions were 
applied to the microscopic elamination of the duplicate samples of the twelve 
cements sent to Professor White. The source of each samples was first disguised 
from the operator in order to eradicate any knowledge of the physical properties of 
the cements until after the predictions from the microscopic examinations had 
been made. Table VIII includes Professor White's diagnosis of these cements, as 
well as our microscopic and sieve tests of fineness. ' 

^^ It is known that manufacturers sometimes test the soundness of their cement 
before the plaster has been added and that this practice has caused the shipment 
of .unsound material. 

* Table VI records phenomena which under certain conditions may impose 
even greater limitations to these interpretations. It imposes a condition, however, 
which is almost negligible in practical cement manufacture, namely: the aeration 
of ground cement in very thin layers. In commercial cements the percentage 
of free lime which will not react with the test solution because of a coating 
of carbonate, is usually insignificant. 



392 



REIBLING AND REYES. 



Table VIII. — Soundness 



6 
o 

1 


1. 


Nature of 
sample. 


Fineness. 


Classifi- 
cation 

by 
White. 


Microscopic examination (duplicate 
samples, Bureau of Science). 


Micro- 
scopic. 




it 
P 


1 


D 


Average sample 
taken from 1 
barrel of ce- 


Medium 


80.2 


94.8 


Poor__ 


An almost immediate formation of 
a mass of long needles (single, 
clusters, and rosettes). Little 






ment. 




1 




additional formation after 5 min- 
utes. No plumes appear in 1 
















hour. 


2 


^ 


do 


___do 


77.4 


93.0 


— do- 


An almost immediate formation of 
many rosettes of needles. Little 




















change after 5 minutes. No 
















plumes appear in 1 hour. 


3 


H 


do 


Fine.-.- 


88.4 


97.2 


Pass- 
able. 


A fair amount of well-defined long 
needles and rosettes. Little 
changes after 5 minutes. One 














feathery crystal appears in 1 hour. 


4 


K 


Average mixture 
of many sam- 


—do— - 


89. 99. 


— do- 


A few defined rosettes and short 
and long needles. No plumes or 






ples taken from 










feathers in 1 hour. Short, 






different bar- 










stubby rosettes of needles develop 






rels. (A quick- 










after 5 minutes. Slides show very 






setting cement.) 










little free lime. 


5 


M 


Average sample 
taken from 1 
barrel of c e - 
ment. 


Medium 


79.8 


94.8 


Poor— 


An almost immediate formation of 
a mass of well-defined long and 
short rosettes of needles. Needles 
continued to grow in number until 
slide presents almost a white ap- 
pearance. No plumes in 1 hour. 


6 


N 


do 


Rather 
coarse. 


76.4 


95.6 


— do- 


Long needle and somewhat feathery 
short needles. Little change after 














5 minutes. No plumes in 1 hour. 


7 





Average mixture 
of many sam- 
ples taken from 
different bar- 
rels. 


Medium 


78.0 


96.0 


— do- 


Long needles and somewhat feath- 
ery short rosettes of needles. A 
greater quantity after 15 minutes. 
Two plumes form in 1 hour. 


8 





Average sample 
taken from 1 
barrel of ce- 
ment. 


Fine.— 


87.4 


97.8 


— do- 


A great mass of rosettes of needles 
(more than any of the above) and 
short. No plumes. 


9 





Unaerated 
ground rotary 
cement receiv- 
ed from cement 
factory. 


Coarse— 


74.8 


88.81 


Bad - 


An almost immediate formation of 
rosettes of needles and plumes in 
numbers. Radiating needles bor- 
der some feathery petals. 



PROPERTIES OF PORTLAND CEMENT. 



393 



versus free lime. 



Interpretation of microscopic test. 


Result of soundness tests. 




Soundness. 










Condition and amount 
of free lime. 






In steam. 


In air. 


In water. 










In steam. 


In air. 


In water. 








No sintered lime. Lit- 


Sound 


Sound 


Sound 


Sound 


Sound 


Sound. 


tle if any nonsintered 














lime. Slaked lime 














in abundance. (A 














well-seasoned c e - 














ment.) 














Very little, if any, sin- 


do — 


do-... 


do..__ 


do 


do 


Do. 


tered lime. A little 














slaked lime. A well- 














seasoned cement. 














Very little sintered 


do.— 


do.... 


do..-. 


do 


....-do.... 


Do. 


lime. Fair amount 














of slaked lime. (A 














well-seasoned hard- 














burned cement.) 














No sintered lime. 


do..._ 


do.-.. 


do—. 


do 


do..._ 


Do. 


Slaked lime and a 














little nonsintered 














lime. 














No sintered lime. 


U n c e r- 


Probably 


Probably 


Off plates, 


do.— 


Do. 


Nonsintered and 


tain. 


sound. 


sound. 


but sound. 






slaked lime in abun- 














dance. (An under- 














burned cement fairly 














well seasoned.) 














A little sintered lime. 


Sound 


Sound 


Sound 


Sound 


Sound 


Sound. 


Slaked lime. (A fair- 














ly well-seasoned ce- 














ment.) 














A small amount of 


do 


do...- 


do...- 


do 


do—. 


Do. 


wintered lime. A lit- 














tle nonsintered lime. 














Slaked lime in abun- 














dance. (Fairly well- 














seasoned under- 














burned cement.) 














No sintered lime. A 


U n c e r- 


U n G e r- 


U n c e r- 


Badly dis- 


Sound oflf 


Sound off 


large amount of 


tain. 


tain. 


tain. 


integrat- 


plate. 


plate. 


slaked and nonsin- 








ed. 






tered lime. (Partial- 














ly seasoned under- 














burned cement.) 














A considerable amount 


Probably 


Probably 


Probably 


Slightly 


do 


Do. 


of sintered lime and 


sound. 


sound. 


sound. 


disinte- 






slakedlime. (A 








. grated. 






slightly seasoned. 














well-burned cement.) 















394 



REIBLING AND REYES. 



Table VIII. — Soundness 









Fineness. 




6 


1 


Nature of 
sample. 


Micro- 
scopic. 


il 
II 


6^ jClassifi- 
S > cation 
-^-i by 
g>^ ! White. 


Microscopic examination (duplicate 
samples, Bureau of Science). 


10 





Same as above 


Coarse— 


75.3 


90.1 'Sound. 


5 minutes, 15 minutes, 1 hour. No 






exposed to 






crystals. 






aeration in pan 


^ 










36 days with 










daily remixing. 






11 





Rotary clinker 
aerated in lab- 
oratory 3 
months and 
then ground in 
ball mill. 


Fine___J 81.6 99.0 Bad __ 

' 1 


A fair amount of long rosettes of 
needles somewhat feathery. No 
increase after 5 minutes. One 
feathery petal in 1 hour. 


12 





Same as No. 10 
except that 1.5 


Coarse __ 


75.8 i 90.1 'Sound. 

1 ! 

1 : 


No plumes and only a few rosettes 
of needles. 






per cent plaster 


; 1 








' was mixed 


' ; ' 








with the un- 


1 










aerated cement. 


i 


i 





It will be seen that the agreement between the microscopic evidence 
and results of the physical test is closer than was to be «ixpected. 



PROPERTIES OF PORTLAND CEMENT. 

versus free lime — Continued. 



395 



Interpretation of microscopic test. 


Result of soundness tests. 


Condition and amount 
of free lime. 


Soundness. 


In steam. 


In air. 


In water. 


In steam. 


In air. 


In water. 


No free lime 

A little sintered lime 
and a little slaked 
lim'e. (A well- 
burned or a well- 
seasoned cement.) 

Only a trace of slaked 
lime. (A very well- 
burned or a very 
well-seasoned c e - 
ment.) 


Sound 

do--„ 

do..__ 


Sound 

do__„ 

do____ 


Sound 

do_... 

do.___ 


Sound 

Small radi- 
al cracks. 
Hard, but 
warped 
^ inch. 

Sound 


Sound 

do.-_. 

do.— 


Sound. 
Do. 

Do. 

1 



396 REIBLING AND REYES. 

Part II. 

THE SEASONING OF PORTLAND (3EMENT. 

The practice of exposing unsound Portland cements to the atmosphere 
until they became sound dates back to the beginning of their manufacture. 
Although certain producers in Europe grind their cement in the presence 
of live steam and others weather their clinker in the open, unprotected 
from the sun and rain, the majority of cements still are seasoned under 
cover by aerating the clinker or the ground material in storage bins. 

Much has been written on the subject of aeration,^^ but the various 
conclusions which have been drawn to explain results so diversified have 
tended rather to promote confusion than certainty. 

Seasoning improves the soundness of unsound cements and therefore 
an almost universal belief seems to exist that seasoning improves all 
Portland cements. 

Important concrete specifications frequently demand a well-seasoned cement. 
For instance, the rigid concrete specifications for the Galveston causeway,** 
require that "all cement must have been seasoned, or subjected to aeration for at 
least thirty days, previous to leaving the mill." 

On the other hand, W. Lawrence Gadd'^ recently conducted experiments from 
which he concluded that "Portland cement made from rotary cement kiln clinker 
does not improve on aeration, but on the contrary becomes more expansive." 

This contradiction alone should convince the consumer that if he places an 
age limit on Portland cement, his reason for so doing should be based upon 
something more tangible than the simple belief that age improves it. 

Although it now generally is conceded that Portland cement properly made 
should stand all standard tests without seasoning as soon as it has lost the heat 
it receives in grinding, some writers, such as R. K. Meade, believe that the engineer 
need not "concern himself whether the manufacturer prefers to make cement which 
is sound when fresh, or whether he prefers to age it sound in his stock house. 
Cement which has seasoned sound is just as good as one which was sound when 
freshly made." 

The effects of aeration on commercial cements were discussed in a 
previous paper from this laboratory ^^ and it was shown that the changes 
produced by aeration were often suflBcient so to alter the tests as to give 
unsatisfactory instead of satisfactory results. It also developed that ad- 
ditional seasoning would pften cause some normal cements to become 
quick setting and others to become more susceptible to change in setting 
time at different temperatures, and that the strength of all perfectly 

2»Candlot, M., Cement et Chaux Hydrauliques, Paris, (1891); Spalding, 
Frederick C, Hydraulic Cement, New York (1904), 4, 56, 80; Taylor & Thompson, 
Concrete, Plain and Reinforced, New York (1907), 62; Meade, Richard K., 
Ghem. Eng. (1907), 5, 341. 

'''' Concrete (1910), 10, 62. 

''Cement Age (1906), 6, 47. 

•^This Journal, Sec, A (1908), 3, 137. 



PROPERTIES OF PORTLAND CEMENT. 



397 



sound cements was lowered by additional seasoning. These changes were 
attributed to the slaking of lime and subsequent reaction with carbon 
dioxide, the lime thereby being rendered an inert substance. It was 
pointed out that only those cements which contained an excess of lime, 
or which were improperly burned, were improved by seasoning. Because 
of these and other considerations the belief was expressed that underburn- 
ing is fatal to the efficiency of Portland cement to be used in the Tropics, 
and exception was taken to the published statements of several author- 
ities on this subject concerning the significance and value of the specific 
gravity and soundness tests. 

As our studies of aerating cements progressed we soon convinced 
ourselves that different methods of seasoning and storing the same 
cement produced different effects upon its physical and chemical prop- 
erties, and that different cements were influenced in unlike manners by 
the same conditions of seasoning. Consequently, we found it necessary 
thoroughly to consider all of these questions in our work. 

THE SEASONING OF GROUND CEMENT. 

The penetration of air into a mass of undisturbed commercial cement 
is shown in Tables IX and X. 

The same experiment was repeated on a larger scale with brand 
cement. 

About 2 kilograms of Portland cement (brand D, Table A) were placed in an 
open, cylindrical glass jar. The cement was not packed nor tamped, but was 
rendered only slightly compact by tapping the sides of the cylinder. In this 
manner it was exposed in the laboratory for one month during the dry season, from 
August 24 to September 24, 1908, at the end of which time samples were taken 
from different depths and the percentage of loss by ignition, moisture, and 
carbon dioxide in each determined. (Table IX.) 

Table IX. — Penetration of air into undisturbed cement. 



! 

Sample number. 


Average dis- 
tance from 
surface 


Loss 
by igni- 
tion.* 


Carbon 
diox- 
ide. 


Mois- 
ture, b 


1 _ 


mm. 

0- 3 

3- 6 

6i- 13 

13i- 26 

25i- 51 

51-127 

127i-203 


Per ct. 
8.96 
7.57 
4.89 
4.44 
4.44 
4.34 
4.34 


Per ct. 
4.86 
3.68 
0.92 
0.72 
0.67 
0.65 
0.66 


Per ct. 
0.97 
0.77 
0.66 
0.65 
0.51 
0.48 
0.44 


2 


3 


4 

5 


6 


7 





« Since only part of the combined water is driven off at a low red heat, the percentage 
of the latter can not be determined from this table. However, Table XII gives us this 
information. 

^ The term "moisture" refers to water which is expelled below 110'. 



398 



REIBLING AND REYES. 



The same experiment was repeated on a larger scale with brand 
cement. 

In this case the cement was exposed in a 10-liter, square, coverless can for 
six months, from September 1, 1909, to January 2, 1910. The can was placed 
near an open window so as to insure an ample change of fresh air. Table X 
records the results. 

Table X. — Penetration of adr into undisturbed cement. 



Sample number. 


Distance 

from 
surface. 


Loss 
by igni- 
tion. 


Carbon- 
diox- 
ide. 


Mois- 
ture. 

Per ct. 
1.75 
l.;^l 
0.86 
0.62 
0.55 


Condition of 
cement. 


1 


mm. 
0- 3 
3- 13 
13- 25 
25- 76 
76-127 


Per ct. 
12.72 
9.28 
4.12 
3.66 
3.60 


Per ct. 
8.56 
5.88 
0.84 
0.53 
''0.35 




2 


Caked hard. 


3 _ 


Caked. 


4 .. 

5_ _ L 


Slightly caked. 
Not caked. 




Do. 



a The unexposed cement contained 0.35 per cent of carbon-dioxide. 

These figures show that only exposed surfaces of ground Portland 
cement are subjected to aeration, and consequently the material under- 
neath this seasoned surface may possess physical characteristics differing 
entirely from those of the upper layers. While the percentage of carbon- 
dioxide changes but very little at a distance greater than 13 millimeters 
below the surface of the cement, water absorbed from the atmosphere 
may slowly penetrate further. This is shown by Table XI. 

Table XI. — Penetration of moisture into undisturbed cement. 



Constituents. 


Cement 
before 
being 

aerated. 

_... 

Per cent. 
0.48 
0.12 
0.00 
0.36 


Aerated 
cement." 


Loss by ignition __ ___ _ 


Per cent. 
1.88 


Carbon dioxide . 


0.12 


Moisture 


0.23 


Water of combination liberated at a low, red heat 


1.52 







» Cement aerated 4 months and 17 days in open glass cylinders ; samples taken 127 
millimeters below the surface. 

The absorption of water and carbon-dioxide from the air by small 
quantities of cement which are frequently remixed, is made clear by 
Table XII. 

The conditions in this instance were the following: Three brands (O, H, and 
M) were used. In each case 60 grams of cement placed in 260 cubic centimeter 
beakers, each of the same diameter, were mixed and weighed daily. An empty 
beaker of the same size collected so little dust (0.002 gram in 2 months) as to 
make a correction for dust unnecessary. The first figures in Table XII give the 
average of three samples. At the intervals stated, one beaker was removed from 
each set and the percentage of volatile constituents in the cement determined. 



PROPERTIES OF PORTLAND CEMENT, 



399 



«3i 






X 



e3 





1^ 


£ 


_ 


2.32 
3.58 

4.98 




1^ 


"IP 


2.01 
1.89 

1.99 



2.51 


i 

W 


ii 


0.48 
1.33 
2.00 

2.01 


Ik 

5'" 


0.61 

3.90 

. 6.94 

9.54 




3.10 
7.13 
10.93 

14.06 


In- 
crease 

in 
weight. 


|lS8t^SS5§^S^8SS^$5^5§^a^ 


lOTHrH(MCOT»iiOCOI>r^COOiOi-HrHW<NeOeO 


1. 






2.21 
2.84 

4.93 


s?i||ii^iii||§8||jj||s 

<N 1 ! 1 1 1 rH 1 1 1 1 1 rH 1 1 , 1 1 t 1 C4 


■H 

S2 


0.85 

1.37 

1.58 


1.86 


ik 

6'°~ 


■ 
1.38 

5.05 

8.58 

11.42 


02 *C 
.Q 


s 


4.75 
8.08 
11.99 

15.90 


^s 


!^S^^?3gBSJ5^:?^^S5S§S^^^^ 




5 S.s.^ 


lOOrHMCOCC-fiOiO^Ot^OOOOOf-trHC^ItNOT 


d 
-d 

c 




S^ 


1.90 
3.54 

4.25 


£ 


2.00 

2.05 
2.30 


II 


0.28 
1.37- 
1.72 

1.83 


Carbon- 
diox- 
ide. 


0.41 


4.36 

7.11 
8.99 




2.69 


7,32 

10.88 
13.12 


In- 
crease 

in 
weight. 


i^?i8^gs?:s^SSS^?3gg8S^ 


idrHC^c^co'^iAdt-ooodoJd.-i-Hf-icl^'eJ 


Number of days 

aerated (Oct., 1908, 

to Apr., 1909). 


Before aeration 

3 

5 

12 

26 

33 

40 

47 

54 

61 , 

75__. 

89-_ 

103 . 

124 

138 

152 

164 

178 

194 



S o 



- fe ^ 


in 


-s &^ 


o 


d fl 


Y- 




^s t 


- 


could 
I low 
off d 
ion. 


o 

OS 



^ CS ,^. 



O «M 
03 S 



O 1-t tH C<l <N « ci 



X, ^ 



5 5 'i 






a g 



s .2 ^ 



Xi 03 



f2 



- 53 O. 



»4 CO 

Ii. 



99667- 



400 



REIBLING AND REYES. 



The increase in different volatile constituents of cement is graph- 
ically shown by figure 12. 



















___— 














,,iA^ 










--- 








/ 


.^^ 
X 

y 


L 


)5SjXi^ 


jjTlOj^U-- 












/ 


/) 














^ 




Q: 




y 




C^ 


oftDiSHS. 










« 




/ . ^ 


















/ 








WATER 


JTR COMB 
IFCOMBINA 


TION LIBERA 


no M AL0\A/ RED HEAT 


r> 


^ 


^g^0^ 






MC 

AGL 


] IN DAY 


s 









Pig. 12. — Diagram showing the characteristic effects of thorough aeration on the 
different volatile constituents of Portland cement. 

The greatest increase is in the percentage of carbon dioxide. The 
figures in Tables IX^ X^ XI and XII confirm the microscopic evidence 
that free lime in Portland cement exposed to the atmosphere changes to 
the carbonate soon after hydrating. 

The microscopic evidence that the change of lime to carbonate tends 
to make the free lime in cement more inert is confirmed by the decrease 
in the rapidity of absorption of carbon dioxide and water and the gradual 
increase in the percentage of free moisture. 

These facts are important by reason of their bearing upon the explana- 
tion of the changes sometimes noted in the physical properties of cements 
which have been stored in air-tight vessels^ because active^ free lime in 
cement so preserved will tend to unite with the moisture which originally 
was present^ and the slaked lime so produced may have a different effect 
upon the physical properties of the cement than a corresponding amount 
of sintered or nonsintered lime. Therefore, it is our experience that 
Portland cements which have been stored while protected from the 
atmosphere seldom contain a high percentage of moisture. Cements that 
do are usually high in other volatile constituents and always show little 



PROPERTIES OP PORTLAND CEMENT. 401 

or no active unslaked lime when the test solution is applied. Hard- 
burned, fresh cements often contain a high percentage of water which can 
be expelled below 110 degrees. If such cements are stored in air-tight 
receptacles it almost invariably happens that the percentage of this 
constituent gradually decreases, but the decrease is extremely slow. We 
can account for the slowness of this change by the fact that the free lime 
in the hard-burned cement is all sintered at a high temperature and 
therefore it hydrates very slowly. One sample of such cement originally 
contained 1.54 per cent of moisture and produced a fair amount of plume- 
like crystals when the test solution was applied. The microscopic test, 
from time to time, showed a gradual decrease in the quantity of plume- 
like crystals and a corresponding increase in long, slender, radiating 
needles. But only after nine months had the hydration of the sintered 
Time become so complete that the microscope failed to show plumes. 
The cement at that time contained 0.42 per cent of moisture. 

Clifford Richardson ^* suggests a different explanation for the changes noted 
in cements which have been stored in air-tight receptacles. However, his theory 
that the tension in solid solutions of calcium silicates and aluminates is released 
by changes in temperature, etc., thus setting free some aluminate, is hardly 
consistent with our knowledge of the behavior of solid solutions; It is unlikely 
that a cement cooled from a white heat, aged, again heated during the grinding 
process, and then slowly cooled in a pulverized state, would undergo any further 
changes in its eutectic curve because of climatic changes in temperature alone. 

Table XII demonstrates that the amount of combined water which is 
driven off during ignition remains fairly constant. All of the moisture 
which combines directly with the lime is driven off at a low red heat. 
The additional water of combination which shows a steady increase in 
Table XII indicates that processes other than hydration of lime have 
taken place, and that some hydraulic cementive action has set in. 

Although the duplicate samples checked very closely, Table XII shows 
that brand I) absorbed the least quantity of volatile constituents. There- 
fore, experiments were conducted to ascertain what effect the nature of 
the cement would have upon the rapidity and the amount of absorption. 

The rate of absorption decreases very rapidly as the reaction proceeds. 
This is apparent from Table XII, but the cements there considered had 
already been subjected to aeration. Tablp XIII records the figures ob- 
tained by aerating a cement which previously had not been exposed to 
the atmosphere. This material was ground from unaerated rotary 
'clinker, and before being exposed in the beakers its loss by ignition was 
only 0.31 per cent. 

" Constitution of Portland Cement, read before the Ass. of Port. Cemt. Mfrs. 
at Atlantic City, June, 1904. 



402 REIBLING AND REYES. 

Table XIII. — The influence of deration on the rapidity of absorption. 



Material. 


Per cent of increase in weight at different ages. 




3 
days. 


5 
days. 


15 
days. 


1 
month. 


2 
months 


3 
months 


4 
months 


5 
months 


6 
months 

8.0 


Hard-burned rotary clinker 
cement (brand 0)._ 


1.3 


1.8 


3.2 


4.6 


6.3 


7.3 


7.6 


7.8 





The influence of a difference in the size of the particles is. shown by 
Table XIV (brand G). The characteristic properties of absorption as 
related to the fine and coarse particles in commercial Portland cement, 
are also illustrated by the figures recorded for brand N. 

Table XIV. — The influence of fineness on the rate and amount of absorption, 



§ 


Fineness. 


i. 


Per cent of increase in weight at different ages. 


i 


i 

'V 
00 








1 


6 

c 
o 
S 

CO 


a 
o 

B 


.El 

1 

S 


1 

B 


G 


As received commercially.. 

Reground to pass a 200- 

mesh sieve 

Fine material passing a 
200-mesh sieve _ _ 


3.29 
3.50 
2.70 
0.56 


0.5 
1.2 
1.5 
0.8 


0.8 
1.8 
2.5 
1.4 


L9 
3.9 
4.6 
3.0 


3.2 
6.0 
6.9 
4.2 


5.4 
9.0 
10.5 
5.5 


7.0 
11.4 
13.1 

6.0 


8.0 
13.3 
15.0 

6.3 


8.8 
15.0 
16.5 

6.4 


9.6 
16.6 
18.0 


N 


Coarse material retained 
on a 200-me8h sieve 


6.4 



The great differences shown by the last column of Table XIV indicate 
that the seasoning of cement is more or less confined to the surface of the 
particles. 

The finer particles of commercial Portland cement usually show a 
higher, and often a very much higher loss by ignition than the coarser 
ones. This may be due to greater surface exposure, as well as to the fact 
that the softest clinker grinds to the finest powder. Consequently, con- 
clusions on the effects of ^e grinding, drawn from experiments in which 
the fine and coarse material used are separated from the same cement, and 
where such conclusions are based upon the physical properties of the 
two cements thus obtained, are apt to be erroneous. -Failure to consider 
the fact that the finer grinding, or regrinding, often increases the per- 
centage of active lime has also caused confusion and error. 

The effect of the degree of burning upon the absortion of moisture and 
carbon-dioxide by Portland cement is a more complicated consideration. 
It is almost an impossibility to grind two cements exactly to the same 
fineness; there also is no accurate method for measuring the degree of 



PROPERTIES OF PORTLAND CEMENT. 



403 



fineness. Consequently^ no conclusions can be drawn from slight dif- 
ferences in results. 

The results recorded in Table XV were gained with cements ground in a 
ball-mill until they gave about the same residue on the 100 and 200-mesh standard 
sieves; the well-burned material required the longer grinding. A microscopic 
examination then indicated little difference in the relative amounts of impalpable 
powder. Each cement was sieved and divided into two parts, that which passed 
the 200-mesh sieve being separated from that which passed the 100, but remained 
on the 200. 



Table XV. — The influence of the degree of burning on the rate of absorption. 
[Figures show percentages.] 



i - 




Fineness. 


6 
o 


Increase 


in weight at different 
ages. 


fa 


s 


Sample. 


1 




.1 

1 








1^ 


1 

o 
1 'O 


i Through a 
200-mesh 
1 sieve. 


28.1 

j 




1 

CO 




1 


i 

CO 


1 

i 


Qa 


Cement ground from un- 










aerated clinker: 








1 














Underburned ce- 




1 


















ment— 






















Coarse particles__ 


0.0 


100 


0.42 


0.19 


0.53 


1.16 


2.10 


3.39 


3.98 


4.40 




Fine particles 


100.0 


100 


0.52 


0.26 


0.62 


1.48 


2.46 


3.52 


6.19 


6.71 


! 


Well-burned cement- 
























Coarse particles__ 


0.0 


100 


0.30 


0.18 


0.44 


1.05 


2.00 


3.16 


3.60 


3.90 




Fine particles_.._: 


100.0 


100 


0.33 


0.35 


0.85 


1.74 


2.83 


4.15 


7.63 


7.98 





Cement ground from 








i 














clinker which was ae- 














i 






rated 3 months in the 




















laboratory: 








; 












Underburned ce- 
























ment — 
























Coarse particles.- 


0.0 


100 






0.49 




1.46 


1.91 


«>2.09 
6.09 






Fine particles 


100.0 


100 


2.20 





0.73 




1.83 


3.25 


8.29 




Well-burned cement- 
























Coarse particles__ 


0.0 


100 






0.69 




1.91 


2.89 


3.21 






Fine particles 


100.0 


100 


0.59 




1.83 




3.14 


4.72 


7.89 


8.48 




Underburned ce- 


84.0 


100 


200 




0.89 




2.09 


3.41 


5.95 


7.95 




ment not sieved, 






















+2. 5 per cent 






















plaster. 












! 




i 





* Similar results were obtained with cement ground from brand L clinker, 
b 64 days old. 

The figures given in Table XV do not show the differences in results we had 
anticipated. Making all allowance for the possible variations in fineness, the 
well-burned cement absorbed carbon dioxide and water as readily as the under- 
burned material. However, when these cements were later subjected to White's 
free lime test, they all demonstrated the presence of unhydrated lime in consider- 
able quantity. Three of the four cements also disintegrated during the steam 
test. 



404 REIBLING AND REYES. 

Table XVI. — Free lime and soundness of the cements recorded in Table XV. 



Sample 
No. 


Description. 


Microscopic test. 


Soundness (5-hour steam 
test). 


Oi 


Nona era ted under- 


Plumes and needles in abundance 


Completely disintegrated. 


1 


burned clinker. 


but not as many plumes as No. 2. 




; («)2 


Nonaerated hard- 


Mostly good large plumes (very 


Very slightly d i s i n t e- 




burned clinker. 


bad). 


grated soft and off plate. 


i 3 


Aerated underburned 


A few plumes, many rosettes of 


Sound. 




clinker. 


needles. Not nearly so bad as 
No. 4. 
Plumes in long needles In abun- 




4 


Ae r a t e d hard-burned 


Soft, slightly disintegrated 




clinker. 


dance (very bad). 


and off plate. 



* These are the same cements given in Tables III and V. 

When the experiment was repeated, using selected, perfectly sintered 

clinker which contained no free lime and which remained perfectly 

sound after the boiling test, the results were more conclusive. (Table 
XVII.) 



Table XVII. — The influence of the degree of fusion on the rapidity of absorption 

{brand, L cement). 

[Figures show percentages.] 



Description. 



Perfectly fused clinker con- 
taining no free lime 

Do 

Perfectly fused clinker con- 
taining no free lime + 2 
per cent plaster 

Underburned clinker con- 
taining free lime 

Do 

Underburned clinker con- 
taining free lime -f 2 per 
cent plaster 



Fineness.* 


'^A 


^Xi 


II i 




100 


100 


100 


100 


100 


100 


90.8 


100 


90.8 


100 


90.8 


100 



Increase in weight at 
different ages. 



0.40 ; 0.30 1.32 
0.40 1 0.38 1.32 



2.00 
2.09 



0.40 


0.38 


0.58 


0.98 


0.53 


0.99 


0.53 


0.80 



2.74 
2.65 



3.29 
3.38 



3.35 



3.60 
3.60 



4.45 
4.55 



S'O 



0.70 
0.68 



0.67 
0.61 



0.84 



9*3 



<a 3 
15 OT crj 



a-s 



2.90 
2.92 



3.94 



3.88 



« The clinker which contained no free lime was ground much finer than the underburned 
clinker In order to insure the accuracy of the main conclusion. Had the cements been 
ground to the same fineness the difference would have been greater, quantitatively. 

When these figures are plotted in curves (figure 13) those for the 
underburned cement result in a parabola while the perfectly fused cement 
gives a straight line. Other absorption curves are also plotted in figure 
13 ; these show the variability which has been encountered. 



PROPERTIES OF PORTLAND CEMENT. 



405 




Fig. 13. — Diagram showing variations in the effects of thorough aeration on the 
volatile constituents of Portland cement. 

Of course, atmospheric conditions exert a marked influence upon the 
amount of water and of carbon dioxide absorbed. In rainy weather, for 
instance, cements frequently contain more than 2 per cent of moisture. 
This tends to accelerate the hydration and to promote setting and caking. 
Therefore, all of the recorded figures and curves are more or less subject 
to variation. However, the information gained from these experiments 
is applicable to practical considerations, as manufacturing conditions 
also involve similar exposures and variations in atmospheric influences. 

THE SEASONING OF PORTLAND CEMENT CLINKERS. 

The action of water and air on Portland cement clinker is radically 
different from that on the ground commercial product. 

Tables XV and XVI show that hard-burned ^^ clinkers may produce 
unsound cements containing a considerable amount of sintered free lime ; 
and furthermore, that prolonged aeration of such clinkers may fail to 
produce soundness, or to slake all of the free lime. On the other hand, 
underburned clinkers from the same kiln and similar raw material acted 
differently. Seasoned during the same time and in the same manner, 
they produced sound cement. This is the opposite to the behavior of 

""^ Well-hurned and hard-hurned designate olinkers. which have undergone 
incipient fusion and which have sintered into dense, hard, rock-like clinkers 
which are very difficult to crush and grind and which do not disintegrate readily. 
When such clinkers contain no free lime, they are regarded as being perfectly 
sintered. 



406 REIBLING AND REYES. 

ground cement which is being aerated (Table VII), but a study of com- 
mercial clinkers gives an explanation for this difference and likewise 
reveals the causes of others. 

The presence of free lime in hard-burned clinker has always been traced 
hy us to too coarse grinding or poor mixing of the raw material, but 
never to separation from a previously combined state. 

The eflfect of coarse grinding on the fusion product was demonstrated by 
Campbell and Ball."® They were unable to burn the raw material secured from 
a Lehigh district plant into a clinker which would produce sound, unseasoned 
cement, even though the temperature of their experimental rotary klin was 
raised as high as 1,612 degrees. When the same raw meal had been ground finer, 
so that 98.0 instead of 86.6 per cent passed a 100-mesh sieve, the sound product 
was obtained at 1,475 degrees. They concluded that the coarse particles of 
calcined limestone had failed to combine and atributed the unsoundness to free 
lime. Later, White" confirmed the correctness of these conclusions. 

Campbell worked with a miniature furnace. Despite the longer time 
during which the raw meal is confined in the clinkering zone of com- 
mercial kilns, our microscopic, physical, and chemical examinations of 
commercial products and investigations at factories, all confirm the 
belief that in the majority of instances the raw meal is too coarse to 
produce perfectly fused clinker. 

For instance, the product of one large Portland cement mill always showed 
free li